The source code of Delphi calling Tessercat 4.1 for OCR text recognition supports direct compilation of almost all versions of DELPHI. I tested Delphi7, Delphi2007, and DelphiXE2 and can be compiled and used normally. Contains all DLL library files and Chinese and English punctuation training fonts (v4.1.0.20190314). The recognition effect is very good. It can normally recognize 99% of Chinese, uppercase and lowercase English characters, punctuation marks, etc., which can basically achieve practical value.
The recognition effect is as shown below (basically it can reach 99.9% accuracy) , and can support various formats of images such as png, jpg, gif, tiff, etc.:
Chinese and English, punctuation marks, upper and lower case, mixed recognition effects:
Pure English recognition effect:
Comparison of AI recognition rate between fuzzy image recognition and the latest QQ screenshots: