(NET) NET (2019 год)

Tesseract OCR Project template

This is my template of OCR projects. Firstly need to add Tesseract package from Nuget. Also you need to download tessdata. By the way, you can download it by unix utility CURL.



After that need to place data to BIN folder and include it to project.



And this is simple test project from documentation for quick start.


   1:  Imports System.Reflection
   2:  Imports Tesseract
   3:   
   4:   
   5:  Module Module1
   6:      Public Sub Main(ByVal args As String())
   7:          Dim testImagePath = "upwork-sample-2-1.png"
   8:   
   9:          Try
  10:              Dim logger = New FormattedConsoleLogger()
  11:              Dim resultPrinter = New ResultPrinter(logger)
  12:   
  13:              Dim path = IO.Path.GetDirectoryName(Assembly.GetExecutingAssembly().CodeBase)
  14:              path = IO.Path.Combine(path, "tessdata")
  15:              path = path.Replace("file:\", "")
  16:   
  17:              Using engine = New TesseractEngine(path, "eng", EngineMode.[Default])
  18:   
  19:                  'engine.SetVariable("tessedit_char_whitelist", "1234567890abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ.,")
  20:                  'engine.SetVariable("tessedit_unrej_any_wd", True)
  21:   
  22:                  Using img = Pix.LoadFromFile(testImagePath)
  23:   
  24:                      Using logger.Begin("Process image " & testImagePath)
  25:                          Dim i = 1
  26:   
  27:                          Using page = engine.Process(img)
  28:                              Dim text = page.GetText()
  29:                              logger.Log("Text: {0}", text)
  30:                              logger.Log("Mean confidence: {0}", page.GetMeanConfidence())
  31:   
  32:                              Using iter = page.GetIterator()
  33:                                  iter.Begin()
  34:   
  35:                                  Do
  36:   
  37:                                      If i Mod 2 = 0 Then
  38:   
  39:                                          Using logger.Begin("Line {0}", i)
  40:   
  41:                                              Do
  42:   
  43:                                                  Using logger.Begin("Word Iteration")
  44:   
  45:                                                      If iter.IsAtBeginningOf(PageIteratorLevel.Block) Then
  46:                                                          logger.Log("New block")
  47:                                                      End If
  48:   
  49:                                                      If iter.IsAtBeginningOf(PageIteratorLevel.Para) Then
  50:                                                          logger.Log("New paragraph")
  51:                                                      End If
  52:   
  53:                                                      If iter.IsAtBeginningOf(PageIteratorLevel.TextLine) Then
  54:                                                          logger.Log("New line")
  55:                                                      End If
  56:   
  57:                                                      logger.Log("word: " & iter.GetText(PageIteratorLevel.Word))
  58:                                                  End Using
  59:                                              Loop While iter.[Next](PageIteratorLevel.TextLine, PageIteratorLevel.Word)
  60:                                          End Using
  61:                                      End If
  62:   
  63:                                      i += 1
  64:                                  Loop While iter.[Next](PageIteratorLevel.Para, PageIteratorLevel.TextLine)
  65:                              End Using
  66:                          End Using
  67:                      End Using
  68:                  End Using
  69:              End Using
  70:   
  71:          Catch ex As Exception
  72:              Trace.TraceError(ex.ToString())
  73:              Console.WriteLine("Unexpected Error: " & ex.Message)
  74:              Console.WriteLine("Details: ")
  75:              Console.WriteLine(ex.ToString())
  76:          End Try
  77:   
  78:          Console.Write("Press any key to continue . . . ")
  79:          Console.ReadKey(True)
  80:      End Sub
  81:   
  82:  End Module

This project use Iterator function and sophisticated formated console logger.


   1:  Imports System
   2:  Imports System.Collections.Generic
   3:  Imports System.Text
   4:  Imports Tesseract
   5:   
   6:   
   7:  Public Class FormattedConsoleLogger
   8:          Const Tab As String = "    "
   9:   
  10:          Private Class Scope
  11:              Inherits DisposableBase
  12:   
  13:              Private indentLevel As Integer
  14:              Private indent As String
  15:              Private container As FormattedConsoleLogger
  16:   
  17:              Public Sub New(ByVal container As FormattedConsoleLogger, ByVal indentLevel As Integer)
  18:                  Me.container = container
  19:                  Me.indentLevel = indentLevel
  20:                  Dim indent As StringBuilder = New StringBuilder()
  21:   
  22:                  For i As Integer = 0 To indentLevel - 1
  23:                      indent.Append(Tab)
  24:                  Next
  25:   
  26:                  Me.indent = indent.ToString()
  27:              End Sub
  28:   
  29:              Public Sub Log(ByVal format As String, ByVal args As Object())
  30:                  Dim message = String.Format(format, args)
  31:                  Dim indentedMessage As StringBuilder = New StringBuilder(message.Length + indent.Length * 10)
  32:                  Dim i As Integer = 0
  33:                  Dim isNewLine As Boolean = True
  34:   
  35:                  While i < message.Length
  36:   
  37:                      If message.Length > i AndAlso message(i) = vbCr AndAlso message(i + 1) = vbLf Then
  38:                          indentedMessage.AppendLine()
  39:                          isNewLine = True
  40:                          i += 2
  41:                      ElseIf message(i) = vbCr OrElse message(i) = vbLf Then
  42:                          indentedMessage.AppendLine()
  43:                          isNewLine = True
  44:                          i += 1
  45:                      Else
  46:   
  47:                          If isNewLine Then
  48:                              indentedMessage.Append(indent)
  49:                              isNewLine = False
  50:                          End If
  51:   
  52:                          indentedMessage.Append(message(i))
  53:                          i += 1
  54:                      End If
  55:                  End While
  56:   
  57:                  Console.WriteLine(indentedMessage.ToString())
  58:              End Sub
  59:   
  60:              Public Function Begin() As Scope
  61:                  Return New Scope(container, indentLevel + 1)
  62:              End Function
  63:   
  64:              Protected Overrides Sub Dispose(ByVal disposing As Boolean)
  65:                  If disposing Then
  66:                      Dim scope = container.scopes.Pop()
  67:   
  68:                      If scope IsNot Me Then
  69:                          Throw New InvalidOperationException("Format scope removed out of order.")
  70:                      End If
  71:                  End If
  72:              End Sub
  73:          End Class
  74:   
  75:          Private scopes As Stack(Of Scope) = New Stack(Of Scope)()
  76:   
  77:          Public Function Begin(ByVal title As String, ParamArray args As Object()) As IDisposable
  78:              Log(title, args)
  79:              Dim scope As Scope
  80:   
  81:              If scopes.Count = 0 Then
  82:                  scope = New Scope(Me, 1)
  83:              Else
  84:                  scope = ActiveScope.Begin()
  85:              End If
  86:   
  87:              scopes.Push(scope)
  88:              Return scope
  89:          End Function
  90:   
  91:          Public Sub Log(ByVal format As String, ParamArray args As Object())
  92:              If scopes.Count > 0 Then
  93:                  ActiveScope.Log(format, args)
  94:              Else
  95:                  Console.WriteLine(String.Format(format, args))
  96:              End If
  97:          End Sub
  98:   
  99:          Private ReadOnly Property ActiveScope As Scope
 100:              Get
 101:                  Dim top = scopes.Peek()
 102:                  If top Is Nothing Then Throw New InvalidOperationException("No current scope")
 103:                  Return top
 104:              End Get
 105:          End Property
 106:      End Class
 107:   

   1:  Imports Tesseract
   2:   
   3:  Public Class ResultPrinter
   4:      ReadOnly logger As FormattedConsoleLogger
   5:   
   6:      Public Sub New(ByVal logger As FormattedConsoleLogger)
   7:          Me.logger = logger
   8:      End Sub
   9:   
  10:      Public Sub Print(ByVal iter As ResultIterator)
  11:          logger.Log("Is beginning of block: {0}", iter.IsAtBeginningOf(PageIteratorLevel.Block))
  12:          logger.Log("Is beginning of para: {0}", iter.IsAtBeginningOf(PageIteratorLevel.Para))
  13:          logger.Log("Is beginning of text line: {0}", iter.IsAtBeginningOf(PageIteratorLevel.TextLine))
  14:          logger.Log("Is beginning of word: {0}", iter.IsAtBeginningOf(PageIteratorLevel.Word))
  15:          logger.Log("Is beginning of symbol: {0}", iter.IsAtBeginningOf(PageIteratorLevel.Symbol))
  16:          logger.Log("Block text: ""{0}""", iter.GetText(PageIteratorLevel.Block))
  17:          logger.Log("Para text: ""{0}""", iter.GetText(PageIteratorLevel.Para))
  18:          logger.Log("TextLine text: ""{0}""", iter.GetText(PageIteratorLevel.TextLine))
  19:          logger.Log("Word text: ""{0}""", iter.GetText(PageIteratorLevel.Word))
  20:          logger.Log("Symbol text: ""{0}""", iter.GetText(PageIteratorLevel.Symbol))
  21:      End Sub
  22:  End Class

But no advantage of using this stupid addition to this project. You may see somple logger for 5 string of code, for example from this page How to intercept exception & console output & debug trace and show it in textbox..



Much bigger problem is 648 parameters to tune OCR engine - Affecting Tesseract OCR engine with special parameters.

And next most biggest problem is quality of image. For example in this image with high quality all text is recognize (but not with yellow background !).



But is you have low quality image (in screen below you can see the same image, but in low quality) we receive unexpected result.





Comments ( )
Link to this page: http://www.vb-net.com/TesseractOCR/index.htm
<На главную>  <В раздел ASP>  <В раздел NET>  <В раздел SQL>  <В раздел Разное>  <Написать автору>  < THANKS ME>