change how OCR checks if a string has proper text

Author: Torwent

Source/simba.pixelocr.pas

@@ -75,8 +75,8 @@   TPixelOCR = record
     function RecognizeLines(Image: TSimbaImage; constref Font: TPixelFont; Bounds: TBox): TStringArray;
   end;
 
-const
-  ALPHA_NUM_SYMBOLS = ['a'..'z', 'A'..'Z', '0'..'9', '%', '&', '#', '$', '[', ']', '{', '}', '@', '!', '?'];
+var
+  ALPHA_NUM_SYMBOLS: array [Char] of Boolean;
 
 implementation
 
@@ -86,6 +86,17 @@ implementation
   simba.vartype_pointarray,
   simba.fs;
 
+initialization
+const
+  CHARACTERS = ['a'..'z', 'A'..'Z', '0'..'9', '%', '&', '#', '$', '[', ']', '{', '}', '@', '!', '?'];
+var
+  c: Char;
+begin
+  FillChar(AlphaNumSymTable, SizeOf(AlphaNumSymTable), False);
+  for c in CHARACTERS do
+    ALPHA_NUM_SYMBOLS[c] := True;
+end;
+
 function IsSimilar(const Image: TSimbaImage; const X, Y: Integer; const Color2: TColorBGRA; const Tol: Single): Boolean; inline;
 const
   MAX_DISTANCE_RGB = Single(441.672955930064); // Sqrt(Sqr(255) + Sqr(255) + Sqr(255))
@@ -112,13 +123,21 @@ function GetGlyph(const Font: PPixelFont; const c: Char): PPixelFontGlyph; inlin
   Result := nil;
 end;
 
-function ContainsAlphaNumSym(const Text: string): Boolean; inline;
+function IsAlphaNumSym(const text: string): Boolean; inline;
 var
-  I: Integer;
+  i, count, needed: Integer;
 begin
-  for I := 1 to Length(Text) do
-    if Text[I] in ALPHA_NUM_SYMBOLS then
-      Exit(True);
+  //50% of text
+  needed := (Length(text) * 50 + 99) div 100;
+
+  for i := 1 to Length(text) do
+    if ALPHA_NUM_SYMBOLS[text[i]] then
+    begin
+      Inc(count);
+      //if 50% of the text is ALPHA_NUM_SYMBOLS we are good.
+      if count >= needed then
+        Exit(True);
+    end;
 
   Result := False;
 end;
@@ -589,7 +608,7 @@ function TPixelOCR.RecognizeLines(Image: TSimbaImage; constref Font: TPixelFont;
       // OCR the row and some extra rows
       Match := Temp._RecognizeXY(Image, @Font, X, Y, Font.MaxGlyphHeight div 2, isBinary);
       // Ensure that actual Text was extracted, not just a symbol mess of short or small character symbols.
-      if ContainsAlphaNumSym(Match.Text) then
+      if IsAlphaNumSym(Match.Text) then
         Result := True;
     end;
   end;