question with screencapture, pixelchecksum

[jsteng]

Hi,

I am experimenting on _ScreenCapture_Capture at the moment and wants to clarify something:

Return Value

Returns a handle to an HBITMAP if $sFileName is empty.

By Handle do you mean like a pointer to the actual data? New to AutoIT, so i am not familiar with the terms used/refered by AutoIT.

If so, how can i access this data? or more importantly, i would like to do a checksum/hash on it, ie to represent that part of the screen into a shorter data that i can easily compare with.

My intention is to do simple pattern recognition/comparison on a fix x1y1-x2y2 window.

A full time OCR would be better but that is slow unfortunately.

Considering i would like to do a checksum/hash, will PixelChecksum do the trick?

ie, can Pixelchecksum be used to represent a particular area's pattern? according to the document, pixelchecksum is intended to check for change in the area but not pattern.

what i want to do:

I want to monitor a spot in the desktop 100x10 pixels (LABEL) in a fix spot.

There are 8 possible patterns on that spot (ie 8 choices in the label)

depending on what the value on the label, an input will have correspondingly diffent set of possible answers.

i would therefore like the script to be able to answer this input field based on the pattern on this LABEL.

currently i am doing an imagesearch by comparing the fix spot LABEL with 8 diffent png files. though it is working, itis not efficient because it can take at most 8 imagesearches to find a match. as it is, it is taking 0.1 seconds to perform imagesearch and 8 of that is 0.8 seconds, average 0.4 seconds.

if somehow I can capture the area into a hash or string or integer, i can easily then do a SELECT CASE search which is extremely faster.

[JohnOne]

PixelChecksum is good for that.

[jsteng]

PixelChecksum is good for that.

I sure like what you said.

will test it once i got back to office.

many thanks

[BrewManNH]

This sounds a lot like a captcha, is that what you're trying to code for?