New OCR Script: Come and get it!

[deo]

First Off, I was having trouble finding an OCR script on here that could be considered reliable. Most of the OCR's I've seen only counts the width of the characters. The downside of that is that with some fonts, different letters will have the same credentials, therefore rendering the script useless.

I wanted an OCR that would be able to differentiate between all characters, not just based off of a X coord, but a Y coord as well.

Second, I was having trouble getting any help on the subject and thought that this community could use something useful and reliable. So I spent the weekend working on this and have decieded to release this to everyone in hopes that it will help someone else.

Here's how it works:

The script will scan from left to right, and from BOTTOM to Top. The bottom-left pixel for each letter will be given a coord of (0,0) and will scan and detect from there.

Example:

Instead of performing multiple loops and checking arrays, this script check's to see what the character is after it's done scanning a single letter.

In order for this script to work, a "baseline" must be specified. This baseline is the very-bottom pixel of ANY letter of the area you are scanning. This would include lowercase "g", "j", "p", "q", "y", etc. Because it scans from bottom to top, the baseline is the start of the scan. The actual width doesnt matter, as long as its wider than the text you want to scan, but the baseline has to be EXACT. The baseline should equal Y position + Height of area to scan.

This is my first script, so the code might not be as advanced as some of the guru's around here, so feel free to give me feedback on what could make this script better/faster.

The downside of this script is that it only has the ability to read one line of text, it doesnt support multi-line text.

Anyway, download it and try it out. Please let me know what you think.

~Deo

DeoOcr.rar

Edited March 11, 2007 by deo

[deo]

I also want to make mention that with $Debug set to 1, and if the parameters are correct, the script will build a definition's list (in the form of a Switch/Case) for the font specified.

[McGod]

Shouldn't this be in example scripts?

[deo]

Hmm, didnt quite think of that. Mod's feel free to move if you wish.

[SmOke_N]

Easy to look through that's for sure... nice work, and I like what you were thinking when you wrote it. I have to say that I'd still like to see a bit different optimization, and arrays in the _CheckChar() function, but all in all, you could actually make this sounder than others for sure.

When I get a bit more time, I'll have a "deeper" look.

Also... Didn't you mean to post this in the scripts and scraps forum? If so, I'm sure a mod will have no problems moving it for you.

Edit:

Ooops, someone already stated the forum... and I even gave it it's old name lol.

Edited March 11, 2007 by SmOke_N

[deo]

When I get a bit more time, I'll have a "deeper" look.

Yes, please do. I'd like to make this as optimized and fast as it can possibly be. Im sure alot of people would appreciate something like this. I sure would have a few days ago

[SmOke_N]

Yes, please do. I'd like to make this as optimized and fast as it can possibly be. Im sure alot of people would appreciate something like this. I sure would have a few days ago

It seems to be font and or size specific. What font and size did you use?

[deo]

Its a FFF font (http://fontsforflash.com/). If you want to give it a different font, you will need to specify the $CharMaxWidth & $CharMaxHeight, and the base line has to be exact for detection to take place. The area that you're scanning needs to be the height of the font itself.

[deo]

Can a mod please move this thread to the example scripts forum? I think it would be more fitting there and I dont want to double post....

[Rizzet]

Hi, im been playing little with your OCR and i like it alot. I put in a little "shade" in it, but havnt been tested enogh. Can you give me a fast curse how the cord system work... i tried to make my own Numbers to it but the OCR dont find it (in the Debug it find the 3 as a B but it doesnt print a B or a 3 )

Edited April 6, 2007 by Rizzet

[Rizzet]

I like your script alot, i have now learned how to use it and im using it in 1 of my scripts. but well with some "fonts" dont know why but it gets messed up and wont work.

Now i found out that its not reseting the Y axel everytime its a new Y line so i made this little change and works fine now

expandcollapse popup

For $tLeft = $Left To (($Width - 1) + $Left)
            $FilledLine = 0
            For $tTop = $BaseLine To $BaseLine - $CharMaxHeight Step -1         
                If Hex(PixelGetColor($tLeft, $tTop), 6) = $CharColor Then
                    If $Debug = 1 Then
                        GUICtrlCreatePic("img\blue.bmp", $tLeft, $tTop+$BaseLine, 1, 1)
                        GUISetState(@SW_SHOW)
                    EndIf
                    ;_msg($X & @CR & $Y)
                    $A &= $X & $Y
                    $FilledLine = 1
                EndIf
                If ($Y < ($Height - 1)) Then
                    $Y += 1
                Else
                    $Y = 0
                EndIf
            Next

$Y = 0 ;<--------------------------------------------------This is the change Now everytime it comes to a new line the Y resets to 0

            If $X <= $CharMaxWidth Then
                ;Check for an empty vertical line
                If $FilledLine = 0 Then
                    If $Debug = 1 Then
                        GUICtrlCreatePic("img\red.bmp", $tLeft, $tTop, 1, 11)
                        GUISetState(@SW_SHOW)
                    EndIf
                
                    If $A <> "" Then
                        ;We found a space between letters, lets check and see what letter we found.
                        $StringReturn &= _CheckChar()
                        $A = ""
                    EndIf
                    $X = 0
                Else
                    $X += 1
                EndIf
            Else
                $X = 0
            EndIf
        Next