Popular Post smbape Posted December 18, 2022 Popular Post Posted December 18, 2022 (edited) After a long hesitation about whether it would be fun to do or not, I finally did it. This UDF provides a way to use mediapipe in AutoIt The usage is similar to the python usage of mediapipe Prerequisites Download and extract autoit-mediapipe-com-1.0.0-opencv-5.0.0-1.7z into a folder Download the opencv UDF from here Sources Here Documentation A generated documentation for functions is available here (v1.0.0-1) Examples More examples can be found here (v1.0.0-1) To run them, please follow these instructions (v1.0.0-1) Face Detection with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/face_detector/python/face_detector.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/face_detector/python/face_detector.ipynb ;~ Title: Face Detection with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" ; STEP 1: Import the necessary modules. Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Main() Func Main() Local $_IMAGE_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\brother-sister-girl-family-boy-977170.jpg" Local $_IMAGE_URL = "https://i.imgur.com/Vu2Nqwb.jpg" Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\blaze_face_short_range.tflite" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/face_detector/blaze_face_short_range/float16/1/blaze_face_short_range.tflite" Local $url, $file_path Local $sample_files[] = [ _ _Mediapipe_Tuple($_IMAGE_FILE, $_IMAGE_URL), _ _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) _ ] For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; STEP 2: Create a FaceDetector object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.FaceDetectorOptions(_Mediapipe_Params("base_options", $base_options)) Local $detector = $vision.FaceDetector.create_from_options($options) ; STEP 3: Load the input image. Local $image = $mp.Image.create_from_file($_IMAGE_FILE) ; Compute the scale to make drawn elements visible when the image is resized for display Local $scale = 1 / resize_and_show($image.mat_view(), Default, False) ; STEP 4: Detect faces in the input image. Local $detection_result = $detector.detect($image) ; STEP 5: Process the detection result. In this case, visualize it. Local $annotated_image = visualize($image.mat_view(), $detection_result, $scale) resize_and_show($annotated_image, "face_detector") $cv.waitKey() EndFunc ;==>Main Func isclose($a, $b) Return Abs($a - $b) <= 1E-6 EndFunc ;==>isclose ; Checks if the float value is between 0 and 1. Func is_valid_normalized_value($value) Return $value >= 0 And $value <= 1 Or isclose(0, $value) Or isclose(1, $value) EndFunc ;==>is_valid_normalized_value #cs Converts normalized value pair to pixel coordinates. #ce Func _normalized_to_pixel_coordinates($normalized_x, $normalized_y, $image_width, $image_height) If Not (is_valid_normalized_value($normalized_x) And is_valid_normalized_value($normalized_y)) Then ; TODO: Draw coordinates even if it's outside of the image bounds. Return Default EndIf Local $x_px = _Min(Floor($normalized_x * $image_width), $image_width - 1) Local $y_px = _Min(Floor($normalized_y * $image_height), $image_height - 1) Return _OpenCV_Point($x_px, $y_px) EndFunc ;==>_normalized_to_pixel_coordinates #cs Draws bounding boxes and keypoints on the input image and return it. Args: image: The input RGB image. detection_result: The list of all "Detection" entities to be visualize. Returns: Image with bounding boxes. #ce Func visualize($rgb_image, $detection_result, $scale = 1.0) Local $MARGIN = 10 * $scale ; pixels Local $ROW_SIZE = 10 ; pixels Local $FONT_SIZE = $scale Local $FONT_THICKNESS = 2 * $scale Local $TEXT_COLOR = _OpenCV_RGB(255, 0, 0) ; red Local $bbox_thickness = 3 * $scale Local $keypoint_color = _OpenCV_RGB(0, 255, 0) Local $keypoint_thickness = 2 * $scale Local $keypoint_radius = 2 * $scale Local $annotated_image = $cv.cvtColor($rgb_image, $CV_COLOR_RGB2BGR) Local $width = $rgb_image.width Local $height = $rgb_image.height Local $bbox, $start_point, $end_point, $keypoint_px Local $category, $category_name, $probability, $result_text, $text_location For $detection In $detection_result.detections ; Draw bounding_box $bbox = $detection.bounding_box $start_point = _OpenCV_Point($bbox.origin_x, $bbox.origin_y) $end_point = _OpenCV_Point($bbox.origin_x + $bbox.width, $bbox.origin_y + $bbox.height) $cv.rectangle($annotated_image, $start_point, $end_point, $TEXT_COLOR, $bbox_thickness) ; Draw keypoints For $keypoint In $detection.keypoints $keypoint_px = _normalized_to_pixel_coordinates($keypoint.x, $keypoint.y, $width, $height) $cv.circle($annotated_image, $keypoint_px, $keypoint_thickness, $keypoint_color, $keypoint_radius) Next ; Draw label and score $category = $detection.categories(0) $category_name = $category.category_name $probability = Round($category.score, 2) $result_text = $category_name & ' (' & $probability & ')' $text_location = _OpenCV_Point($MARGIN + $bbox.origin_x, $MARGIN + $ROW_SIZE + $bbox.origin_y) $cv.putText($annotated_image, $result_text, $text_location, $CV_FONT_HERSHEY_PLAIN, $FONT_SIZE, $TEXT_COLOR, $FONT_THICKNESS) Next Return $annotated_image EndFunc ;==>visualize Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Face Landmarks Detection with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/face_landmarker/python/%5BMediaPipe_Python_Tasks%5D_Face_Landmarker.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/face_landmarker/python/%5BMediaPipe_Python_Tasks%5D_Face_Landmarker.ipynb ;~ Title: Face Landmarks Detection with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" ; STEP 1: Import the necessary modules. Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Global $drawing_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision.drawing_utils") _AssertIsObj($drawing_utils, "Failed to load mediapipe.tasks.autoit.vision.drawing_utils") Global $drawing_styles = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision.drawing_styles") _AssertIsObj($drawing_styles, "Failed to load mediapipe.tasks.autoit.vision.drawing_styles") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Main() Func Main() Local $_IMAGE_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\business-person.png" Local $_IMAGE_URL = "https://storage.googleapis.com/mediapipe-assets/business-person.png" Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\face_landmarker.task" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/face_landmarker/face_landmarker/float16/1/face_landmarker.task" Local $url, $file_path Local $sample_files[] = [ _ _Mediapipe_Tuple($_IMAGE_FILE, $_IMAGE_URL), _ _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) _ ] For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; STEP 2: Create a FaceLandmarker object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.FaceLandmarkerOptions(_Mediapipe_Params("base_options", $base_options, _ "output_face_blendshapes", True, _ "output_facial_transformation_matrixes", True, _ "num_faces", 1)) Local $detector = $vision.FaceLandmarker.create_from_options($options) ; STEP 3: Load the input image. Local $image = $mp.Image.create_from_file($_IMAGE_FILE) ; STEP 4: Detect hand landmarks from the input image. Local $detection_result = $detector.detect($image) ; STEP 5: Process the classification result. In this case, visualize it. Local $annotated_image = draw_landmarks_on_image($image.mat_view(), $detection_result) resize_and_show($annotated_image) $cv.waitKey() EndFunc ;==>Main Func draw_landmarks_on_image($image, $detection_result) ; Compute the scale to make drawn elements visible when the image is resized for display Local $scale = 1 / resize_and_show($image, Default, False) Local $face_landmarks_list = $detection_result.face_landmarks Local $annotated_image = $cv.cvtColor($image, $CV_COLOR_RGB2BGR) ; Loop through the detected faces to visualize. For $face_landmarks In $face_landmarks_list ; Draw the face landmarks. $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_TESSELATION, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_tesselation_style($scale))) $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_CONTOURS, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_contours_style(1, $scale))) $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_LEFT_IRIS, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_iris_connections_style($scale))) $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_RIGHT_IRIS, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_iris_connections_style($scale))) Next Return $annotated_image EndFunc ;==>draw_landmarks_on_image Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Gesture Recognizer with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/gesture_recognizer/python/gesture_recognizer.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/gesture_recognizer/python/gesture_recognizer.ipynb ;~ Title: Gesture Recognizer with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Global $mp_hands = $mp.tasks.vision.HandLandmarksConnections Global $mp_drawing = $mp.tasks.vision.drawing_utils Global $mp_drawing_styles = $mp.tasks.vision.drawing_styles Main() Func Main() Local $IMAGE_FILENAMES[] = ['thumbs_down.jpg', 'victory.jpg', 'thumbs_up.jpg', 'pointing_up.jpg'] Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\gesture_recognizer.task" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/gesture_recognizer/gesture_recognizer/float16/1/gesture_recognizer.task" Local $sample_files[UBound($IMAGE_FILENAMES) + 1] $sample_files[0] = _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) Local $url, $file_path, $name For $i = 0 To UBound($IMAGE_FILENAMES) - 1 $name = $IMAGE_FILENAMES[$i] $file_path = $MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $name $url = "https://storage.googleapis.com/mediapipe-tasks/gesture_recognizer/" & $name $sample_files[$i + 1] = _Mediapipe_Tuple($file_path, $url) Next For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; STEP 2: Create a GestureRecognizer object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.GestureRecognizerOptions(_Mediapipe_Params("base_options", $base_options)) Local $recognizer = $vision.GestureRecognizer.create_from_options($options) Local $image, $recognition_result, $top_gesture, $hands_landmarks For $image_file_name In $IMAGE_FILENAMES ; STEP 3: Load the input image. $image = $mp.Image.create_from_file($MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $image_file_name) ; STEP 4: Recognize gestures in the input image. $recognition_result = $recognizer.recognize($image) ; STEP 5: Process the result. In this case, visualize it. $top_gesture = $recognition_result.gestures(0) (0) $hands_landmarks = $recognition_result.hand_landmarks display_image_with_gestures_and_hand_landmarks($image, $top_gesture, $hands_landmarks) Next $cv.waitKey() EndFunc ;==>Main #cs Displays an image with the gesture category and its score along with the hand landmarks. #ce Func display_image_with_gestures_and_hand_landmarks($image, $gesture, $hands_landmarks) ; Display gestures and hand landmarks. Local $annotated_image = $cv.cvtColor($image.mat_view(), $CV_COLOR_RGB2BGR) Local $title = StringFormat("%s (%.2f)", $gesture.category_name, $gesture.score) ; Compute the scale to make drawn elements visible when the image is resized for display Local $scale = 1 / resize_and_show($annotated_image, Default, False) For $hand_landmarks In $hands_landmarks $mp_drawing.draw_landmarks( _ $annotated_image, _ $hand_landmarks, _ $mp_hands.HAND_CONNECTIONS, _ $mp_drawing_styles.get_default_hand_landmarks_style($scale), _ $mp_drawing_styles.get_default_hand_connections_style($scale)) Next resize_and_show($annotated_image, $title) EndFunc ;==>display_image_with_gestures_and_hand_landmarks Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Hand Landmarks Detection with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/hand_landmarker/python/hand_landmarker.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/hand_landmarker/python/hand_landmarker.ipynb ;~ Title: Hand Landmarks Detection with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Global $mp_hands = $mp.tasks.vision.HandLandmarksConnections Global $mp_drawing = $mp.tasks.vision.drawing_utils Global $mp_drawing_styles = $mp.tasks.vision.drawing_styles Main() Func Main() Local $_IMAGE_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\woman_hands.jpg" Local $_IMAGE_URL = "https://storage.googleapis.com/mediapipe-tasks/hand_landmarker/woman_hands.jpg" Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\hand_landmarker.task" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/hand_landmarker/hand_landmarker/float16/1/hand_landmarker.task" Local $url, $file_path Local $sample_files[] = [ _ _Mediapipe_Tuple($_IMAGE_FILE, $_IMAGE_URL), _ _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) _ ] For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; STEP 2: Create an HandLandmarker object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.HandLandmarkerOptions(_Mediapipe_Params("base_options", $base_options, _ "num_hands", 2)) Local $detector = $vision.HandLandmarker.create_from_options($options) ; STEP 3: Load the input image. Local $image = $mp.Image.create_from_file($_IMAGE_FILE) ; STEP 4: Detect hand landmarks from the input image. Local $detection_result = $detector.detect($image) ; STEP 5: Process the detection result. In this case, visualize it. Local $annotated_image = draw_landmarks_on_image($image.mat_view(), $detection_result) resize_and_show($annotated_image, "hand_landmarker") $cv.waitKey() EndFunc ;==>Main Func draw_landmarks_on_image($rgb_image, $detection_result) ; Compute the scale to make drawn elements visible when the image is resized for display Local $scale = 1 / resize_and_show($rgb_image, Default, False) Local $MARGIN = 10 * $scale ; pixels Local $FONT_SIZE = $scale Local $FONT_THICKNESS = 2 * $scale Local $HANDEDNESS_TEXT_COLOR = _OpenCV_Scalar(88, 205, 54) ; vibrant green Local $hand_landmarks_list = $detection_result.hand_landmarks Local $handedness_list = $detection_result.handedness Local $annotated_image = $cv.cvtColor($rgb_image, $CV_COLOR_RGB2BGR) Local $width = $annotated_image.width Local $height = $annotated_image.height Local $hand_landmarks, $handedness Local $min_x, $min_y, $text_x, $text_y ; Loop through the detected hands to visualize. For $idx = 0 To UBound($hand_landmarks_list) - 1 $hand_landmarks = $hand_landmarks_list[$idx] $handedness = $handedness_list[$idx] $min_x = 1 $min_y = 1 For $landmark In $hand_landmarks If $landmark.x < $min_x Then $min_x = $landmark.x If $landmark.y < $min_y Then $min_y = $landmark.y Next ; Draw the hand landmarks. $mp_drawing.draw_landmarks( _ $annotated_image, _ $hand_landmarks, _ $mp_hands.HAND_CONNECTIONS, _ $mp_drawing_styles.get_default_hand_landmarks_style($scale), _ $mp_drawing_styles.get_default_hand_connections_style($scale)) ; Get the top left corner of the detected hand's bounding box. $text_x = $min_x * $width $text_y = $min_y * $height - $MARGIN ; Draw handedness (left or right hand) on the image. $cv.putText($annotated_image, $handedness[0].category_name, _ _OpenCV_Point($text_x, $text_y), $CV_FONT_HERSHEY_DUPLEX, _ $FONT_SIZE, $HANDEDNESS_TEXT_COLOR, $FONT_THICKNESS, $CV_LINE_AA) Next Return $annotated_image EndFunc ;==>draw_landmarks_on_image Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "image" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Holistic Landmarks Detection with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/7f3cf17410db91f8e084a46f7517df15bce3479a/examples/holistic_landmarker/python/[MediaPipe_Python_Tasks]_Holistic_Landmarker.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/7f3cf17410db91f8e084a46f7517df15bce3479a/examples/holistic_landmarker/python/[MediaPipe_Python_Tasks]_Holistic_Landmarker.ipynb ;~ Title: Holistic Landmarks Detection with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Global $drawing_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision.drawing_utils") _AssertIsObj($drawing_utils, "Failed to load mediapipe.tasks.autoit.vision.drawing_utils") Global $drawing_styles = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision.drawing_styles") _AssertIsObj($drawing_styles, "Failed to load mediapipe.tasks.autoit.vision.drawing_styles") Main() Func Main() Local $_IMAGE_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\business-person.png" Local $_IMAGE_URL = "https://storage.googleapis.com/mediapipe-assets/business-person.png" Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\holistic_landmarker.task" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/holistic_landmarker/holistic_landmarker/float16/1/holistic_landmarker.task" Local $url, $file_path Local $sample_files[] = [ _ _Mediapipe_Tuple($_IMAGE_FILE, $_IMAGE_URL), _ _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) _ ] For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; STEP 2: Create an HolisticLandmarker object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.HolisticLandmarkerOptions(_Mediapipe_Params( _ "base_options", $base_options, _ "output_face_blendshapes", True, _ "output_segmentation_mask", True)) Local $detector = $vision.HolisticLandmarker.create_from_options($options) ; STEP 3: Load the input image. Local $image = $mp.Image.create_from_file($_IMAGE_FILE) ; STEP 4: Detect holistic landmarks from the input image. Local $detection_result = $detector.detect($image) ; STEP 5: Process the detection result. In this case, visualize it. Local $annotated_image = draw_landmarks_on_image($image.mat_view(), $detection_result) resize_and_show($annotated_image, "holistic_landmarker") $cv.waitKey() EndFunc ;==>Main Func draw_landmarks_on_image($rgb_image, $detection_result) ; Compute the scale to make drawn elements visible when the image is resized for display Local $scale = 1 / resize_and_show($rgb_image, Default, False) Local $pose_landmarks = $detection_result.pose_landmarks Local $face_landmarks = $detection_result.face_landmarks Local $left_hand_landmarks = $detection_result.left_hand_landmarks Local $right_hand_landmarks = $detection_result.right_hand_landmarks Local $annotated_image = $cv.cvtColor($rgb_image, $CV_COLOR_RGB2BGR) ; Draw pose landmarks. Local $pose_landmark_style, $pose_connection_style If $pose_landmarks.size() <> 0 Then $pose_landmark_style = $drawing_styles.get_default_pose_landmarks_style($scale) $pose_connection_style = $drawing_utils.DrawingSpec(_Mediapipe_Params("color", _OpenCV_Scalar(0, 255, 0), "thickness", 2 * $scale)) $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $pose_landmarks, _ "connections", $vision.PoseLandmarksConnections.POSE_LANDMARKS, _ "landmark_drawing_spec", $pose_landmark_style, _ "connection_drawing_spec", $pose_connection_style)) EndIf ; Draw face landmarks. If $face_landmarks.size() <> 0 Then $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_TESSELATION, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_tesselation_style(_Mediapipe_Params("scale", $scale)))) $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_CONTOURS, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_contours_style(_Mediapipe_Params("scale", $scale)))) $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_LEFT_IRIS, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_iris_connections_style(_Mediapipe_Params("scale", $scale)))) $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $face_landmarks, _ "connections", $vision.FaceLandmarksConnections.FACE_LANDMARKS_RIGHT_IRIS, _ "landmark_drawing_spec", Null, _ "connection_drawing_spec", $drawing_styles.get_default_face_mesh_iris_connections_style(_Mediapipe_Params("scale", $scale)))) EndIf ; Draw hand landmarks. If $left_hand_landmarks.size() <> 0 Then $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $left_hand_landmarks, _ "connections", $vision.HandLandmarksConnections.HAND_CONNECTIONS, _ "landmark_drawing_spec", $drawing_styles.get_default_hand_landmarks_style(_Mediapipe_Params("scale", $scale)), _ "connection_drawing_spec", $drawing_styles.get_default_hand_connections_style(_Mediapipe_Params("scale", $scale)))) EndIf If $right_hand_landmarks.size() <> 0 Then $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $right_hand_landmarks, _ "connections", $vision.HandLandmarksConnections.HAND_CONNECTIONS, _ "landmark_drawing_spec", $drawing_styles.get_default_hand_landmarks_style(_Mediapipe_Params("scale", $scale)), _ "connection_drawing_spec", $drawing_styles.get_default_hand_connections_style(_Mediapipe_Params("scale", $scale)))) EndIf Return $annotated_image EndFunc ;==>draw_landmarks_on_image Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "image" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Image Classifier with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/image_classification/python/image_classifier.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/image_classification/python/image_classifier.ipynb ;~ Title: Image Classifier with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Main() Func Main() Local $IMAGE_FILENAMES[] = ['burger.jpg', 'cat.jpg'] Local $url, $file_path For $name In $IMAGE_FILENAMES $file_path = $MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $name $url = "https://storage.googleapis.com/mediapipe-tasks/image_classifier/" & $name If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\efficientnet_lite0.tflite" If Not FileExists($_MODEL_FILE) Then $download_utils.download("https://storage.googleapis.com/mediapipe-models/image_classifier/efficientnet_lite0/float32/1/efficientnet_lite0.tflite", $_MODEL_FILE) EndIf ; STEP 2: Create an ImageClassifier object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.ImageClassifierOptions(_Mediapipe_Params("base_options", $base_options, "max_results", 4)) Local $classifier = $vision.ImageClassifier.create_from_options($options) Local $image, $classification_result, $top_category, $title For $image_name In $IMAGE_FILENAMES ; STEP 3: Load the input image. $image = $mp.Image.create_from_file($MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $image_name) ; STEP 4: Classify the input image. $classification_result = $classifier.classify($image) ; STEP 5: Process the classification result. In this case, visualize it. $top_category = $classification_result.classifications(0).categories(0) $title = StringFormat("%s (%.2f)", $top_category.category_name, $top_category.score) resize_and_show($cv.cvtColor($image.mat_view(), $CV_COLOR_RGB2BGR), $title) Next $cv.waitKey() EndFunc ;==>Main Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Image Embedding with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/image_embedder/python/image_embedder.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/image_embedder/python/image_embedder.ipynb ;~ Title: Image Embedding with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Global $cosine_similarity = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.components.utils.cosine_similarity") _AssertIsObj($cosine_similarity, "Failed to load mediapipe.tasks.autoit.components.utils.cosine_similarity") Main() Func Main() Local $IMAGE_FILENAMES[] = ['burger.jpg', 'burger_crop.jpg'] Local $url, $file_path For $name In $IMAGE_FILENAMES $file_path = $MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $name $url = "https://storage.googleapis.com/mediapipe-assets/" & $name If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\mobilenet_v3_small.tflite" If Not FileExists($_MODEL_FILE) Then $download_utils.download("https://storage.googleapis.com/mediapipe-models/image_embedder/mobilenet_v3_small/float32/1/mobilenet_v3_small.tflite", $_MODEL_FILE) EndIf ; Create options for Image Embedder Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $l2_normalize = True ;@param {type:"boolean"} Local $quantize = True ;@param {type:"boolean"} Local $options = $vision.ImageEmbedderOptions(_Mediapipe_Params( _ "base_options", $base_options, _ "l2_normalize", $l2_normalize, _ "quantize", $quantize)) ; Create Image Embedder Local $embedder = $vision.ImageEmbedder.create_from_options($options) ; Format images for MediaPipe Local $first_image = $mp.Image.create_from_file($MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $IMAGE_FILENAMES[0]) Local $second_image = $mp.Image.create_from_file($MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $IMAGE_FILENAMES[1]) Local $first_embedding_result = $embedder.embed($first_image) Local $second_embedding_result = $embedder.embed($second_image) Local $similarity = $cosine_similarity.cosine_similarity($first_embedding_result.embeddings(0), $second_embedding_result.embeddings(0)) ConsoleWrite('@@ Debug(' & @ScriptLineNumber & ') : $similarity = ' & $similarity & @CRLF) ;### Debug Console resize_and_show($cv.cvtColor($first_image.mat_view(), $CV_COLOR_RGB2BGR), $IMAGE_FILENAMES[0]) resize_and_show($cv.cvtColor($second_image.mat_view(), $CV_COLOR_RGB2BGR), $IMAGE_FILENAMES[1]) $cv.waitKey() EndFunc ;==>Main Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Image Segmenter expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/image_segmentation/python/image_segmentation.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/image_segmentation/python/image_segmentation.ipynb ;~ Title: Image Segmenter #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Main() Func Main() Local $IMAGE_FILENAMES[] = ['segmentation_input_rotation0.jpg'] Local $url, $file_path For $name In $IMAGE_FILENAMES $file_path = $MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $name $url = "https://storage.googleapis.com/mediapipe-assets/" & $name If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\deeplab_v3.tflite" If Not FileExists($_MODEL_FILE) Then $download_utils.download("https://storage.googleapis.com/mediapipe-models/image_segmenter/deeplab_v3/float32/1/deeplab_v3.tflite", $_MODEL_FILE) EndIf Local $BG_COLOR = _OpenCV_Scalar(192, 192, 192) ; gray Local $FG_COLOR = _OpenCV_Scalar(255, 255, 255) ; white ; Create the options that will be used for ImageSegmenter Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.ImageSegmenterOptions(_Mediapipe_Params("base_options", $base_options, _ "output_category_mask", True)) ; Create the image segmenter Local $segmenter = $vision.ImageSegmenter.create_from_options($options) Local $image, $segmentation_result, $category_mask, $image_data Local $fg_image, $bg_image, $fg_mask Local $output_image, $blurred_image ; Loop through demo image(s) For $image_file_name In $IMAGE_FILENAMES ; Create the MediaPipe image file that will be segmented $image = $mp.Image.create_from_file($MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $image_file_name) ; Retrieve the masks for the segmented image $segmentation_result = $segmenter.segment($image) $category_mask = $segmentation_result.category_mask ; mediapipe uses RGB images while opencv uses BGR images $image_data = $cv.cvtColor($image.mat_view(), $CV_COLOR_RGB2BGR) ; Generate solid color images for showing the output segmentation mask. $fg_image = $cv.Mat.create($image_data.size(), $CV_8UC3, $FG_COLOR) $bg_image = $cv.Mat.create($image_data.size(), $CV_8UC3, $BG_COLOR) ; The foreground mask corresponds to all 'i' pixels where category_mask[i] > 0.2 $fg_mask = $cv.compare($category_mask.mat_view(), 0.2, $CV_CMP_GT) ; Draw fg_image on bg_image only where fg_mask should apply $output_image = $bg_image.copy() $fg_image.copyTo($fg_mask, $output_image) resize_and_show($output_image, 'Segmentation mask of ' & $image_file_name) ; Blur the image only where fg_mask should not apply $blurred_image = $cv.GaussianBlur($image_data, _OpenCV_Size(55, 55), 0) $image_data.copyTo($fg_mask, $blurred_image) resize_and_show($blurred_image, 'Blurred background of ' & $image_file_name) Next $cv.waitKey() EndFunc ;==>Main Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Interactive Image Segmenter expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/interactive_segmentation/python/interactive_segmenter.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/interactive_segmentation/python/interactive_segmenter.ipynb ;~ Title: Interactive Image Segmenter #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Global $containers = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.components.containers") _AssertIsObj($containers, "Failed to load mediapipe.tasks.autoit.components.containers") Main() Func Main() Local $IMAGE_FILENAMES[] = ['cats_and_dogs.jpg'] Local $url, $file_path For $name In $IMAGE_FILENAMES $file_path = $MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $name $url = "https://storage.googleapis.com/mediapipe-assets/" & $name If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\magic_touch.tflite" If Not FileExists($_MODEL_FILE) Then $download_utils.download("https://storage.googleapis.com/mediapipe-models/interactive_segmenter/magic_touch/float32/1/magic_touch.tflite", $_MODEL_FILE) EndIf Local $x = 0.68 ;@param {type:"slider", min:0, max:1, step:0.01} Local $y = 0.68 ;@param {type:"slider", min:0, max:1, step:0.01} Local $BG_COLOR = _OpenCV_Scalar(192, 192, 192) ; gray Local $FG_COLOR = _OpenCV_Scalar(255, 255, 255) ; white Local $OVERLAY_COLOR = _OpenCV_Scalar(100, 100, 0) ; cyan Local $RegionOfInterest = $vision.InteractiveSegmenterLegacyRegionOfInterest Local $NormalizedKeypoint = $containers.keypoint.NormalizedKeypoint ; Create the options that will be used for InteractiveSegmenterLegacy Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.InteractiveSegmenterLegacyOptions(_Mediapipe_Params("base_options", $base_options, _ "output_category_mask", True)) ; Create the interactive segmenter Local $segmenter = $vision.InteractiveSegmenterLegacy.create_from_options($options) Local $image, $roi, $segmentation_result, $category_mask, $image_data Local $fg_image, $bg_image, $fg_mask Local $output_image, $blurred_image, $overlayed_image Local $keypoint_px, $alpha Local $color = _OpenCV_Scalar(255, 255, 0) Local $thickness, $radius, $scale ; Loop through demo image(s) For $image_file_name In $IMAGE_FILENAMES ; Create the MediaPipe image file that will be segmented $image = $mp.Image.create_from_file($MEDIAPIPE_SAMPLES_DATA_PATH & "\" & $image_file_name) ; Retrieve the masks for the segmented image $roi = $RegionOfInterest(_Mediapipe_Params("format", $MEDIAPIPE_TASKS_VISION_INTERACTIVE_SEGMENTER_LEGACY_REGION_OF_INTEREST_FORMAT_KEYPOINT, _ "keypoint", $NormalizedKeypoint($x, $y))) $segmentation_result = $segmenter.segment($image, $roi) $category_mask = $segmentation_result.category_mask ; mediapipe uses RGB images while opencv uses BGR images $image_data = $cv.cvtColor($image.mat_view(), $CV_COLOR_RGB2BGR) ; Generate solid color images for showing the output segmentation mask. $fg_image = $cv.Mat.create($image_data.size(), $CV_8UC3, $FG_COLOR) $bg_image = $cv.Mat.create($image_data.size(), $CV_8UC3, $BG_COLOR) ; The foreground mask corresponds to all 'i' pixels where category_mask[i] > 0.2 $fg_mask = $cv.compare($category_mask.mat_view(), 0.1, $CV_CMP_GT) ; Draw fg_image on bg_image only where fg_mask should apply $output_image = $bg_image.copy() $fg_image.copyTo($fg_mask, $output_image) ; Compute the point of interest coordinates $keypoint_px = _normalized_to_pixel_coordinates($x, $y, $image.width, $image.height) ; Compute the scale to make drawn elements visible when the image is resized for display $scale = 1 / resize_and_show($image, Default, False) $thickness = 10 * $scale $radius = 2 * $scale ; Draw a circle to denote the point of interest $cv.circle($output_image, $keypoint_px, $thickness, $color, $radius) ; Display the segmented image resize_and_show($output_image, 'Segmentation mask of ' & $image_file_name) ; Blur the image only where fg_mask should not apply $blurred_image = $cv.GaussianBlur($image_data, _OpenCV_Size(55, 55), 0) $image_data.copyTo($fg_mask, $blurred_image) ; Draw a circle to denote the point of interest $cv.circle($blurred_image, $keypoint_px, $thickness, $color, $radius) ; Display the blurred image resize_and_show($blurred_image, 'Blurred background of ' & $image_file_name) ; Create an overlay image with the desired color (e.g., (255, 0, 0) for red) $overlayed_image = $cv.Mat.create($image_data.size(), $CV_8UC3, $OVERLAY_COLOR) ; Create an alpha channel based on the segmentation mask with the desired opacity (e.g., 0.7 for 70%) ; fg_mask values are 0 where the mask should not apply and 255 where it should ; multiplying by 0.7 / 255.0 gives values that are 0 where the mask should not apply and 0.7 where it should $alpha = $fg_mask.convertTo($CV_32F, Null, 0.7 / 255.0) ; repeat the alpha mask for each image channel color $alpha = $cv.merge(_OpenCV_Tuple($alpha, $alpha, $alpha)) ; Blend the original image and the overlay image based on the alpha channel $overlayed_image = $cv.add($cv.multiply($image_data, $cv.subtract(1.0, $alpha), Null, Default, $CV_32F), $cv.multiply($overlayed_image, $alpha, Null, Default, $CV_32F)) ; Draw a circle to denote the point of interest $cv.circle($overlayed_image, $keypoint_px, $thickness, $color, $radius) ; Display the overlayed image resize_and_show($overlayed_image, 'Overlayed foreground of ' & $image_file_name) Next $cv.waitKey() EndFunc ;==>Main Func isclose($a, $b) Return Abs($a - $b) <= 1E-6 EndFunc ;==>isclose ; Checks if the float value is between 0 and 1. Func is_valid_normalized_value($value) Return ($value > 0 Or isclose(0, $value)) And ($value < 1 Or isclose(1, $value)) EndFunc ;==>is_valid_normalized_value #cs Converts normalized value pair to pixel coordinates. #ce Func _normalized_to_pixel_coordinates($normalized_x, $normalized_y, $image_width, $image_height) If Not (is_valid_normalized_value($normalized_x) And is_valid_normalized_value($normalized_y)) Then ; TODO: Draw coordinates even if it's outside of the image bounds. Return Default EndIf Local $x_px = _Min(Floor($normalized_x * $image_width), $image_width - 1) Local $y_px = _Min(Floor($normalized_y * $image_height), $image_height - 1) Return _OpenCV_Point($x_px, $y_px) EndFunc ;==>_normalized_to_pixel_coordinates Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Language Detector with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/language_detector/python/%5BMediaPipe_Python_Tasks%5D_Language_Detector.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/language_detector/python/%5BMediaPipe_Python_Tasks%5D_Language_Detector.ipynb ;~ Title: Language Detector with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" ; STEP 1: Import the necessary modules. Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $text = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.text") _AssertIsObj($text, "Failed to load mediapipe.tasks.autoit.text") Main() Func Main() Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\language_detector.tflite" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/language_detector/language_detector/float32/latest/language_detector.tflite" Local $url, $file_path Local $sample_files[] = [ _ _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) _ ] For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; Define the input text that you wants the model to classify. Local $INPUT_TEXT = "分久必合合久必分" ;@param {type:"string"} ; STEP 2: Create a LanguageDetector object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $text.LanguageDetectorOptions(_Mediapipe_Params("base_options", $base_options)) Local $detector = $text.LanguageDetector.create_from_options($options) ; STEP 3: Get the language detcetion result for the input text. Local $detection_result = $detector.detect($INPUT_TEXT) ; STEP 4: Process the detection result and print the languages detected and their scores. For $detection In $detection_result.detections ConsoleWrite(StringFormat("%s: (%.2f)", $detection.language_code, $detection.probability) & @CRLF) Next EndFunc ;==>Main Func _OnAutoItExit() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Object Detection with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/object_detection/python/object_detector.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/object_detection/python/object_detector.ipynb ;~ Title: Object Detection with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Main() Func Main() Local $_IMAGE_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\cat_and_dog.jpg" Local $_IMAGE_URL = "https://storage.googleapis.com/mediapipe-tasks/object_detector/cat_and_dog.jpg" Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\efficientdet_lite0.tflite" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/object_detector/efficientdet_lite0/int8/1/efficientdet_lite0.tflite" Local $url, $file_path Local $sample_files[] = [ _ _Mediapipe_Tuple($_IMAGE_FILE, $_IMAGE_URL), _ _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) _ ] For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; STEP 2: Create an ObjectDetector object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.ObjectDetectorOptions(_Mediapipe_Params("base_options", $base_options, _ "score_threshold", 0.5)) Local $detector = $vision.ObjectDetector.create_from_options($options) ; STEP 3: Load the input image. Local $image = $mp.Image.create_from_file($_IMAGE_FILE) ; Compute the scale to make drawn elements visible when the image is resized for display Local $scale = 1 / resize_and_show($image.mat_view(), Default, False) ; STEP 4: Detect objects in the input image. Local $detection_result = $detector.detect($image) ; STEP 5: Process the detection result. In this case, visualize it. Local $annotated_image = visualize($image.mat_view(), $detection_result, $scale) resize_and_show($annotated_image, "object_detection") $cv.waitKey() ; STEP 6: Closes the detector explicitly when the detector is not used ina context. $detector.close() EndFunc ;==>Main #cs Draws bounding boxes and keypoints on the input image and return it. Args: rgb_image: The input RGB image. detection_result: The list of all "Detection" entities to be visualize. scale: Scale to keep drawing visible after resize Returns: Image with bounding boxes. #ce Func visualize($rgb_image, $detection_result, $scale = 1.0) Local $MARGIN = 10 * $scale ; pixels Local $ROW_SIZE = 10 ; pixels Local $FONT_SIZE = $scale Local $FONT_THICKNESS = $scale Local $TEXT_COLOR = _OpenCV_RGB(255, 0, 0) ; red Local $annotated_image = $cv.cvtColor($rgb_image, $CV_COLOR_RGB2BGR) Local $bbox, $start_point, $end_point Local $category, $category_name, $probability, $result_text, $text_location For $detection In $detection_result.detections ; Draw bounding_box $bbox = $detection.bounding_box $start_point = _OpenCV_Point($bbox.origin_x, $bbox.origin_y) $end_point = _OpenCV_Point($bbox.origin_x + $bbox.width, $bbox.origin_y + $bbox.height) $cv.rectangle($annotated_image, $start_point, $end_point, $TEXT_COLOR, 3) ; Draw label and score $category = $detection.categories(0) $category_name = $category.category_name $probability = Round($category.score, 2) $result_text = $category_name & ' (' & $probability & ')' $text_location = _OpenCV_Point($MARGIN + $bbox.origin_x, $MARGIN + $ROW_SIZE + $bbox.origin_y) $cv.putText($annotated_image, $result_text, $text_location, $CV_FONT_HERSHEY_PLAIN, $FONT_SIZE, $TEXT_COLOR, $FONT_THICKNESS) Next Return $annotated_image EndFunc ;==>visualize Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Pose Landmarks Detection with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/pose_landmarker/python/%5BMediaPipe_Python_Tasks%5D_Pose_Landmarker.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/pose_landmarker/python/%5BMediaPipe_Python_Tasks%5D_Pose_Landmarker.ipynb ;~ Title: Pose Landmarks Detection with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") _OpenCV_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-opencv-com\autoit_opencv_com500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $cv = _OpenCV_get() _AssertIsObj($cv, "Failed to load opencv") Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $drawing_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision.drawing_utils") _AssertIsObj($drawing_utils, "Failed to load mediapipe.tasks.autoit.vision.drawing_utils") Global $drawing_styles = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision.drawing_styles") _AssertIsObj($drawing_styles, "Failed to load mediapipe.tasks.autoit.vision.drawing_styles") Global $vision = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.vision") _AssertIsObj($vision, "Failed to load mediapipe.tasks.autoit.vision") Main() Func Main() Local $_IMAGE_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\girl-4051811_960_720.jpg" Local $_IMAGE_URL = "https://cdn.pixabay.com/photo/2019/03/12/20/39/girl-4051811_960_720.jpg" Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\pose_landmarker_heavy.task" Local $_MODEL_URL = "https://storage.googleapis.com/mediapipe-models/pose_landmarker/pose_landmarker_heavy/float16/1/pose_landmarker_heavy.task" Local $url, $file_path Local $sample_files[] = [ _ _Mediapipe_Tuple($_IMAGE_FILE, $_IMAGE_URL), _ _Mediapipe_Tuple($_MODEL_FILE, $_MODEL_URL) _ ] For $config In $sample_files $file_path = $config[0] $url = $config[1] If Not FileExists($file_path) Then $download_utils.download($url, $file_path) EndIf Next ; STEP 2: Create a PoseLandmarker object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $vision.PoseLandmarkerOptions(_Mediapipe_Params( _ "base_options", $base_options, _ "output_segmentation_masks", True)) Local $detector = $vision.PoseLandmarker.create_from_options($options) ; STEP 3: Load the input image. Local $image = $mp.Image.create_from_file($_IMAGE_FILE) ; STEP 4: Detect pose landmarks from the input image. Local $detection_result = $detector.detect($image) ; STEP 5: Process the detection result. In this case, visualize it. Local $annotated_image = draw_landmarks_on_image($image.mat_view(), $detection_result) ; Display the image resize_and_show($annotated_image, "Pose Landmarks Detection with MediaPipe Tasks : Image") ; Visualize the pose segmentation mask. Local $segmentation_mask = $detection_result.segmentation_masks(0).mat_view() resize_and_show($segmentation_mask, "Pose Landmarks Detection with MediaPipe Tasks : Mask") $cv.waitKey() EndFunc ;==>Main Func draw_landmarks_on_image($rgb_image, $detection_result) ; Compute the scale to make drawn elements visible when the image is resized for display Local $scale = 1 / resize_and_show($rgb_image, Default, False) Local $pose_landmarks_list = $detection_result.pose_landmarks Local $annotated_image = $cv.cvtColor($rgb_image, $CV_COLOR_RGB2BGR) Local $pose_landmark_style = $drawing_styles.get_default_pose_landmarks_style($scale) Local $pose_connection_style = $drawing_utils.DrawingSpec(_Mediapipe_Params("color", _Mediapipe_Tuple(0, 255, 0), "thickness", 2)) ; Loop through the detected poses to visualize. For $pose_landmarks In $pose_landmarks_list ; Draw the pose landmarks. $drawing_utils.draw_landmarks(_Mediapipe_Params( _ "image", $annotated_image, _ "landmark_list", $pose_landmarks, _ "connections", $vision.PoseLandmarksConnections.POSE_LANDMARKS, _ "landmark_drawing_spec", $pose_landmark_style, _ "connection_drawing_spec", $pose_connection_style)) Next Return $annotated_image EndFunc ;==>draw_landmarks_on_image Func resize_and_show($image, $title = Default, $show = Default) If $title == Default Then $title = "" If $show == Default Then $show = True Local Const $DESIRED_HEIGHT = 480 Local Const $DESIRED_WIDTH = 480 Local $w = $image.width Local $h = $image.height If $h < $w Then $h = $h / ($w / $DESIRED_WIDTH) $w = $DESIRED_WIDTH Else $w = $w / ($h / $DESIRED_HEIGHT) $h = $DESIRED_HEIGHT EndIf Local $interpolation = ($DESIRED_WIDTH > $image.width Or $DESIRED_HEIGHT > $image.height) ? $CV_INTER_CUBIC : $CV_INTER_AREA If $show Then Local $img = $cv.resize($image, _OpenCV_Size($w, $h), _OpenCV_Params("interpolation", $interpolation)) $cv.imshow($title, $img.convertToShow()) EndIf Return $w / $image.width EndFunc ;==>resize_and_show Func _OnAutoItExit() _OpenCV_Close() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Text Classifier with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/text_classification/python/text_classifier.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/text_classification/python/text_classifier.ipynb ;~ Title: Text Classifier with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $text = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.text") _AssertIsObj($text, "Failed to load mediapipe.tasks.autoit.text") Main() Func Main() Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\bert_classifier.tflite" If Not FileExists($_MODEL_FILE) Then $download_utils.download("https://storage.googleapis.com/mediapipe-models/text_classifier/bert_classifier/float32/1/bert_classifier.tflite", $_MODEL_FILE) EndIf ; Define the input text that you want the model to classify. Local $INPUT_TEXT = "I'm looking forward to what will come next." ; STEP 2: Create a TextClassifier object. Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) Local $options = $text.TextClassifierOptions(_Mediapipe_Params("base_options", $base_options)) Local $classifier = $text.TextClassifier.create_from_options($options) ; STEP 3: Classify the input text. Local $classification_result = $classifier.classify($INPUT_TEXT) ; STEP 4: Process the classification result. In this case, print out the most likely category. Local $top_category = $classification_result.classifications(0).categories(0) ConsoleWrite('@@ Debug(' & @ScriptLineNumber & ') : ' & StringFormat('%s (%.2f)', $top_category.category_name, $top_category.score) & @CRLF) ;### Debug Console EndFunc ;==>Main Func _OnAutoItExit() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Text Embedding with MediaPipe Tasks expandcollapse popup#Region ;**** Directives created by AutoIt3Wrapper_GUI **** #AutoIt3Wrapper_UseX64=y #AutoIt3Wrapper_Change2CUI=y #AutoIt3Wrapper_Au3Check_Parameters=-d -w 1 -w 2 -w 3 -w 4 -w 5 -w 6 #AutoIt3Wrapper_AU3Check_Stop_OnWarning=y #EndRegion ;**** Directives created by AutoIt3Wrapper_GUI **** ;~ Sources: ;~ https://colab.research.google.com/github/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/text_embedder/python/text_embedder.ipynb ;~ https://github.com/google-ai-edge/mediapipe-samples/blob/3d23f0e459907af064c3e7494dbb180851e1694c/examples/text_embedder/python/text_embedder.ipynb ;~ Title: Text Embedding with MediaPipe Tasks #include "autoit-mediapipe-com\udf\mediapipe_udf_utils.au3" #include "autoit-opencv-com\udf\opencv_udf_utils.au3" _Mediapipe_Open("opencv-5.0.0-windows\opencv\build\x64\vc16\bin\opencv_world500.dll", "autoit-mediapipe-com\autoit_mediapipe_com-1.0.0-500.dll") OnAutoItExitRegister("_OnAutoItExit") ; Where to download data files Global Const $MEDIAPIPE_SAMPLES_DATA_PATH = @ScriptDir & "\examples\data" Global $download_utils = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.core.download_utils") _AssertIsObj($download_utils, "Failed to load mediapipe.tasks.autoit.core.download_utils") ; STEP 1: Import the necessary modules. Global $mp = _Mediapipe_get() _AssertIsObj($mp, "Failed to load mediapipe") Global $autoit = _Mediapipe_ObjCreate("mediapipe.tasks.autoit") _AssertIsObj($autoit, "Failed to load mediapipe.tasks.autoit") Global $text = _Mediapipe_ObjCreate("mediapipe.tasks.autoit.text") _AssertIsObj($text, "Failed to load mediapipe.tasks.autoit.text") Main() Func Main() Local $_MODEL_FILE = $MEDIAPIPE_SAMPLES_DATA_PATH & "\bert_embedder.tflite" If Not FileExists($_MODEL_FILE) Then $download_utils.download("https://storage.googleapis.com/mediapipe-models/text_embedder/bert_embedder/float32/1/bert_embedder.tflite", $_MODEL_FILE) EndIf ; Create your base options with the model that was downloaded earlier Local $base_options = $autoit.BaseOptions(_Mediapipe_Params("model_asset_path", $_MODEL_FILE)) ; Set your values for using normalization and quantization Local $l2_normalize = True ;@param {type:"boolean"} Local $quantize = False ;@param {type:"boolean"} ; Create the final set of options for the Embedder Local $options = $text.TextEmbedderOptions(_Mediapipe_Params( _ "base_options", $base_options, "l2_normalize", $l2_normalize, "quantize", $quantize)) Local $embedder = $text.TextEmbedder.create_from_options($options) ; Retrieve the first and second sets of text that will be compared Local $first_text = "I'm feeling so good" ;@param {type:"string"} Local $second_text = "I'm okay I guess" ;@param {type:"string"} ; Convert both sets of text to embeddings Local $first_embedding_result = $embedder.embed($first_text) Local $second_embedding_result = $embedder.embed($second_text) ; Retrieve the cosine similarity value from both sets of text, then take the ; cosine of that value to receie a decimal similarity value. Local $similarity = $text.TextEmbedder.cosine_similarity($first_embedding_result.embeddings(0), _ $second_embedding_result.embeddings(0)) ConsoleWrite('@@ Debug(' & @ScriptLineNumber & ') : $similarity = ' & $similarity & @CRLF) ;### Debug Console EndFunc ;==>Main Func _OnAutoItExit() _Mediapipe_Close() EndFunc ;==>_OnAutoItExit Func _AssertIsObj($vVal, $sMsg) If Not IsObj($vVal) Then ConsoleWriteError($sMsg & @CRLF) Exit 0x7FFFFFFF EndIf EndFunc ;==>_AssertIsObj Edited Monday at 05:02 AM by smbape Update to OpenCV 5.0.0, Update Mediapie to 1.0.0 malcev, Parsix, Danyfirex and 2 others 4 1
smbape Posted December 18, 2022 Author Posted December 18, 2022 Hi @malcev The UDF of mediapipe is out. Do you happen to know of any mediapipe python examples that are different from what is available at MediaPipe in Python? Everything I come across is a variation of those examples.
malcev Posted December 18, 2022 Posted December 18, 2022 Hi smbape! Sorry, but I dont know any other resources. Only some video tutorials like this: Thank You very much for such great work!
smbape Posted February 19, 2023 Author Posted February 19, 2023 Update mediapipe to 0.9.1 COM registration is no longer required malcev 1
malcev Posted March 5, 2023 Posted March 5, 2023 (edited) Hi, smbape! Will You plan to add object detection? https://developers.google.com/mediapipe/solutions/vision/object_detector/python Also here is new features like audio classifier, gesture recognizer, text classification... https://github.com/googlesamples/mediapipe/tree/main/examples Edited March 5, 2023 by malcev
smbape Posted March 6, 2023 Author Posted March 6, 2023 Hi @malcev, 11 hours ago, malcev said: Will You plan to add object detection? Yes. I was not add because I didn't see it. Thanks for the information.
smbape Posted March 19, 2023 Author Posted March 19, 2023 Hi @malcev After spending a few days on these new features, the code seems unfinished. Some parts have not been tested and there is duplicated code. Even the given example https://github.com/googlesamples/mediapipe/blob/main/examples/text_classification/python/text_classifier.ipynb does not work. Quote Attention: This MediaPipe Solutions Preview is an early release. The MediaPipe Solutions Preview is an early release that is subject to the following limitations: it may have limited support, changes may not be compatible with other pre-general availability versions, and availability may change without notice. Although I have already spent a lot of time on this topic, I will stop for now. At least until these features become stable.
smbape Posted April 28, 2023 Author Posted April 28, 2023 Hi @malcev All python google samples examples, except audio classifier, are available here https://github.com/smbape/node-autoit-mediapipe-com/tree/v0.4.0/examples/googlesamples/examples malcev 1
malcev Posted April 28, 2023 Posted April 28, 2023 Hi, smbape! Very interesting to test it. Thank You very much! Will You plan to add audio classifier example in future?
smbape Posted April 29, 2023 Author Posted April 29, 2023 Hi @malcev Yes I will, when they will have a working c++ example.
smbape Posted June 30, 2024 Author Posted June 30, 2024 (edited) Update MediaPipe to 0.10.14 Edited June 8, 2025 by smbape
smbape Posted August 4, 2025 Author Posted August 4, 2025 Update MediaPipe to 0.10.26 argumentum and Danyfirex 2
smbape Posted Monday at 05:03 AM Author Posted Monday at 05:03 AM Update Mediapie to 1.0.0 argumentum and KaFu 2
water Posted Tuesday at 09:02 AM Posted Tuesday at 09:02 AM Added your UDF to the Wiki. argumentum and smbape 2 My UDFs and Tutorials: Spoiler UDFs: Active Directory (NEW 2024-07-28 - Version 1.6.3.0) - Download - General Help & Support - Example Scripts - Wiki ExcelChart (2017-07-21 - Version 0.4.0.1) - Download - General Help & Support - Example Scripts OutlookEX (2021-11-16 - Version 1.7.0.0) - Download - General Help & Support - Example Scripts - Wiki OutlookEX_GUI (2021-04-13 - Version 1.4.0.0) - Download Outlook Tools (2019-07-22 - Version 0.6.0.0) - Download - General Help & Support - Wiki PowerPoint (2021-08-31 - Version 1.5.0.0) - Download - General Help & Support - Example Scripts - Wiki Task Scheduler (2022-07-28 - Version 1.6.0.1) - Download - General Help & Support - Wiki Standard UDFs: Excel - Example Scripts - Wiki Word - Wiki Tutorials: ADO - Wiki WebDriver - Wiki
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now