UWPOCR - Windows Platform Optical character recognition API Implementation

[Danyfirex]

Hello guys.  I recently saw some posts that Windows 10 provides OCR API. So I decided to create a UDF.

 

 

What's UWPOCR?

UWPOCR UDF is a simple library to use Universal Windows Platform Optical character recognition API.

Features.

• Get Text From Image File.
• Get Text From GDI+ Bitmap.
• Easy to use.

 Usage:

#include "..\UWPOCR.au3"

_Example()

Func _Example()
    Local $sOCRTextResult = _UWPOCR_GetText(FileOpenDialog("Select Image", @ScriptDir & "\", "Images (*.jpg;*.bmp;*.png;*.tif;*.gif)"))
    MsgBox(0,"",$sOCRTextResult)
EndFunc

Get Words Rect(Example):

 

More examples here.

 

Check UWPOCR UDF on GitHub.

 

Saludos

Edited February 4, 2022 by Danyfirex
Update link

[mLipok]

Awesome.

[mLipok]

As so far _UWPOCR_GetText() support only $hBitmap or FileFullPath, could you add a feature to provide BinaryVariant ?

[Danyfirex]

@mLipok  I think it's not needed you just need to use _GDIPlus_BitmapCreateFromMemory

 

Saludos

Edited January 10, 2022 by Danyfirex

[mLipok]

How to get data from desired/specified window ?

As so far I try:

#include <ScreenCapture.au3>
#include <GDIPlus.au3>
#include "..\UWPOCR.au3"


_Example()

Func _Example()

    #Region - create example bitmap - take a screenshot
    _GDIPlus_Startup()
    ;hImage/hBitmap GDI
    Local $hTimer = TimerInit()
    Local $hHBitmap = _ScreenCapture_CaptureWnd("", WinWait("[REGEXPTITLE:(?i)(.*SciTE.*)]"))
    Local $hBitmap = _GDIPlus_BitmapCreateFromHBITMAP($hHBitmap)
    #EndRegion - create example bitmap  - take a screenshot

    #Region - do the OCR Stuff
    ;Get OCR Text From hImage/hBitmap
;~  _UWPOCR_Log(__UWPOCR_Log) ;Enable Log
    Local $sOCRTextResult = _UWPOCR_GetText($hBitmap, Default, True)
    MsgBox(0, "Time Elapsed: " & TimerDiff($hTimer), $sOCRTextResult)
    #EndRegion - do the OCR Stuff

    #Region - bitmap clean up
    _WinAPI_DeleteObject($hHBitmap)
    _GDIPlus_BitmapDispose($hBitmap)
    _GDIPlus_Shutdown()
    #EndRegion - bitmap clean up

EndFunc   ;==>_Example

But this not works properly.

Edited January 10, 2022 by mLipok

[Danyfirex]

Try with [CLASS:SciTEWindow]

 

Saludos

[mLipok]

So I checked:
 

#include <ScreenCapture.au3>
#include <GDIPlus.au3>
#include "..\UWPOCR.au3"


_Example()

Func _Example()

    #Region - create example bitmap - take a screenshot
    _GDIPlus_Startup()
    ;hImage/hBitmap GDI
    Local $hTimer = TimerInit()
;~  Local $hHBitmap = _ScreenCapture_CaptureWnd("", WinWait("[REGEXPTITLE:(?i)(.*SciTE.*)]"))
;~  Local $hHBitmap = _ScreenCapture_CaptureWnd("", WinWait("[CLASS:SciTEWindow]"))
;~  Local $hHBitmap = _ScreenCapture_CaptureWnd("", WinWait("[CLASS:AVL_AVView]"))
    Local $hHBitmap = _ScreenCapture_CaptureWnd("", WinWait("[CLASS:AcrobatSDIWindow]"))
    Local $hBitmap = _GDIPlus_BitmapCreateFromHBITMAP($hHBitmap)
    #EndRegion - create example bitmap  - take a screenshot

    #Region - do the OCR Stuff
    ;Get OCR Text From hImage/hBitmap
;~  _UWPOCR_Log(__UWPOCR_Log) ;Enable Log
    Local $sOCRTextResult = _UWPOCR_GetText($hBitmap, Default, True)
    MsgBox(0, "Time Elapsed: " & TimerDiff($hTimer), $sOCRTextResult)
    #EndRegion - do the OCR Stuff

    #Region - bitmap clean up
    _WinAPI_DeleteObject($hHBitmap)
    _GDIPlus_BitmapDispose($hBitmap)
    _GDIPlus_Shutdown()
    #EndRegion - bitmap clean up

EndFunc   ;==>_Example

and:

  Quote

Developing for ÂcIcbe Reader - ÂcIcbe Âcrcbat Reader DC (32-bit)
Plik Edycja Widok
1
Podpis
Okno
Pomoc
Developing for Ado.. x
Zaloguj się
Introduction
This chapter describes the supwrted platforms for development using the Acrobat SDK, and summarizes
the technologies available within the SDK.
Supported platforms
This section describes the requirements for using the interapplication communication and Acrobat core
and extendui APIS on all platforms. The follcw.•łng platforms are currently supported for
development with Adobe Rea der:
Solaris•
HP-UX
Details for each platform are described belwa.
Windows
Microsoft Windows 2CKH) with Service Pack 2
Microsoft Windows XP Professional or Home Edition
Development environments
Microsoft Visual Studio AET 2CH)2
Microsoft Visual Studio 2CH)5
Note: For more information on the Windcws environment, su the and Developing P,'ug-ins and
Apm'łcatbns.
Mac OS
Mac OS X versions 10.28 or Later.
Development environment
Note: For more information on Mac OS development environments, see the Ovewiew and Devenp,ing
P,'ug-ins and Applications

 

Expand  

and it is taken from:

[mLipok]

The issues with OCR are becuase I had zoomed out PDF (68%)

[mLipok]

When you get some spare time you can check: https://github.com/DanysysTeam/UWPOCR/pulls

I did some PR

In the next steps, I will try to provide functions such as:

_UWPOCR_GetTextFromFile()
_UWPOCR_GetTextFromScreen()
_UWPOCR_GetTextFromWindow()
_UWPOCR_GetTextFromMemory()

 

Edited January 10, 2022 by mLipok

[jguinch]

@Danyfirex very impressive !

 

[Werty]

This is cool, no more tesseract needed

[drrehak]

is the advantage that we dont need to install tesseract?  is it faster? or is it better in any other way?

Edited January 11, 2022 by drrehak

[Danyfirex]

@drrehak It seems to be it's supported since Windows 10. Faster or better. I don't know. It need to be tested to know it.

 

Saludos

[mLipok]

  On 1/11/2022 at 2:16 PM, drrehak said:

is the advantage that we dont need to install tesseract?

Expand  

Yes. Only Windows 10 + AutoIt

  On 1/11/2022 at 2:16 PM, drrehak said:

is it faster?

Expand  

Please do some test with tesseract and show how you test them (post here: 1. code snippet  2. testing image 3. your timing results).
Then we prepare example for the same source data with this UDF an you will be able to compare.

  On 1/11/2022 at 2:16 PM, drrehak said:

is it better in any other way?

Expand  

As so far nobody asked this question and I doupt that any body use this features with AutoIt any time before , so we can only guess.

btw. I saw that @Danyfirex answered when I was in the middle ofr my writing here.

Edited January 11, 2022 by mLipok

[Melba23]

Danyfirex,

Nice work Dany!!

M23

[Danyfirex]

Here is a script to compare UWPOCR vs Tesseract. (it just load image process and show result in editboxes and display the processed image to compare visually. 

expandcollapse popup

#include <WinAPISys.au3>
#include <WinAPIsysinfoConstants.au3>
#include <StaticConstants.au3>
#include <EditConstants.au3>
#include <WindowsConstants.au3>
#include <GDIPlus.au3>
#include <GUIConstantsEx.au3>
#include <FileConstants.au3>
#include "..\..\UWPOCR.au3"

Global $g_sTesseractPath = @ScriptDir & "\tesseract-ocr-w64-setup-v5.0.0-alpha.20210811\tesseract.exe" ;change me

_TestUWPOCRvsTesseractOCR()

Func _TestUWPOCRvsTesseractOCR()

    Local $sFileOpenDialog = FileOpenDialog("Select images to process", @ScriptDir & "\", "Images (*.jpg;*.png;*.gif;*.tif;*.bmp)", BitOR($FD_FILEMUSTEXIST, $FD_MULTISELECT))
    Local $aFiles = StringSplit($sFileOpenDialog, "|", 1)
    Local $sImagePath = ""
    For $i = 2 To $aFiles[0]
        $sImagePath = $aFiles[$i]
        _GUIComparation($sImagePath, _UWPOCR_GetText($sImagePath), _GetTesseractOCRText($sImagePath))
    Next
EndFunc   ;==>_TestUWPOCRvsTesseractOCR

Func _GUIImageProcessed($sTitle, $sImagePath)
    _GDIPlus_Startup()
    Local Const $iW = @DesktopWidth, $iH = @DesktopHeight

    Local $hImage = _GDIPlus_ImageLoadFromFile($sImagePath) ;create a GDI bitmap by capturing full screen of the desktop

    Local $iWImage = _GDIPlus_ImageGetWidth($hImage)
    Local $iHImage = _GDIPlus_ImageGetHeight($hImage)
    If $iWImage > $iW Or $iHImage > $iH Then
        $iWImage = Int($iWImage / 1.5)
        $iHImage = Int($iHImage / 1.5)
    EndIf

    Local $hBitmap_Scaled = _GDIPlus_ImageResize($hImage, $iWImage, $iHImage) ;resize image

    Local $hGUI = GUICreate($sTitle, $iWImage, $iHImage, -1, -1) ;create a test gui to display the resized image
    GUISetState(@SW_SHOW)

    Local $hGraphics = _GDIPlus_GraphicsCreateFromHWND($hGUI) ;create a graphics object from a window handle
    _GDIPlus_GraphicsDrawImage($hGraphics, $hBitmap_Scaled, 0, 0) ;display scaled image

    ;cleanup resources
    _GDIPlus_GraphicsDispose($hGraphics)
    _GDIPlus_ImageDispose($hImage)
    _GDIPlus_BitmapDispose($hBitmap_Scaled)
    _GDIPlus_Shutdown()
    Return $hGUI
EndFunc   ;==>_GUIImageProcessed

Func _GUIComparation($sTitle, $sUWPOCRText, $sTesseractText)
    Local $hGUIImage = _GUIImageProcessed($sTitle, $sTitle)
    GUIRegisterMsg($WM_NCHITTEST, _WM_NCHITTEST)
    Local $iWidth = @DesktopWidth - 100
    Local $iHeight = @DesktopHeight - 100
    Local $hWindow = GUICreate($sTitle, $iWidth, $iHeight)
    GUISetBkColor(0x000000)
    Local $idtxtUWPOCR = GUICtrlCreateEdit($sUWPOCRText, 0, 30, ($iWidth / 2) - 5, $iHeight - 30, BitOR($ES_WANTRETURN, $WS_VSCROLL, $ES_AUTOVSCROLL))
    GUICtrlSetFont(-1, 11)
    Local $idtxtTesseract = GUICtrlCreateEdit($sTesseractText, ($iWidth / 2) + 5, 30, ($iWidth / 2) - 5, $iHeight - 30, BitOR($ES_WANTRETURN, $WS_VSCROLL, $ES_AUTOVSCROLL))
    GUICtrlSetFont(-1, 11)
    GUICtrlCreateLabel("UWPOCR Output", 1, 5, ($iWidth / 2) - 5, 25, $SS_CENTER)
    GUICtrlSetColor(-1, 0xffffff)
    GUICtrlSetFont(-1, 15)
    GUICtrlCreateLabel("Tesseract Output", ($iWidth / 2) + 5, 5, ($iWidth / 2) - 5, 25, $SS_CENTER)
    GUICtrlSetColor(-1, 0xffffff)
    GUICtrlSetFont(-1, 15)
    GUISetState(@SW_SHOW, $hWindow)
    WinActivate($hGUIImage)
    WinMove($hGUIImage, "", Default, $iHeight / 2)
    MsgBox(0, "Info", "Click to continue...")
    GUIDelete($hWindow)
    GUIDelete($hGUIImage)
EndFunc   ;==>_GUIComparation

Func _WM_NCHITTEST($hWnd, $uMsg, $wParam, $lParam)
    Return $HTCAPTION
EndFunc   ;==>_WM_NCHITTEST


Func _GetTesseractOCRText($sImagePath)
    Local $sFileOutText = @ScriptDir & "\output"
    RunWait('"' & $g_sTesseractPath & '" "' & $sImagePath & '" "' & $sFileOutText & '"', "", @SW_HIDE)
    Local $sOutText = FileRead($sFileOutText & ".txt")
    FileDelete($sOutText)
    Return $sOutText
EndFunc   ;==>_GetTesseractOCRText

Saludos

[SeanDoe]

Hi Danyfirex,

great work! Thank you very much!

Btw:

This:

  On 1/10/2022 at 10:10 PM, mLipok said:

When you get some spare time you can check: https://github.com/DanysysTeam/UWPOCR/pulls

I did some PR

In the next steps, I will try to provide functions such as:

_UWPOCR_GetTextFromFile()
_UWPOCR_GetTextFromScreen()
_UWPOCR_GetTextFromWindow()
_UWPOCR_GetTextFromMemory()

 

Expand  

would be realy awesome too!

Sean

[dmob]

A New Year's gift

Thank you very much @Danyfirex

Saludos 😊

[KaFu]

Check available languages with _ArrayDisplay(_UWPOCR_GetSupportedLanguages())

Install additional language packs like described here:

https://support.microsoft.com/en-us/windows/language-packs-for-windows-a5094319-a92d-18de-5b53-1cfc697cfca8

Select language codes as displayed in above mentioned check:

e.g.  : Local $sOCRTextResult = _UWPOCR_GetText($sImageFilePath,"de-DE",True)

Edited January 12, 2022 by KaFu

[Danyfirex]

Thanks @KaFu I forgot to add that example. 

 

Saludos