/ppocr

Basic OCR.

Primary LanguageC++GNU Lesser General Public License v2.1LGPL-2.1

LGPLv2.1

Learning (./learning)

Generate resources/glyphs/* and resources/datas/* with

$ ./learning "$bindir"/write_glyphs "$bindir"/normalize_glyphs "$bindir"/write_datas

The executables are compiled with bjam. bindir depends on the compiler used.

$ bjam or $ bjam write_glyphs normalize_glyphs write_datas

resources/fonts/{latin,cyrillic}/

./letters and ./generated/

letters contains glyphs of generated/*/*.pnm (generated by font2png (from font2png.cpp) then converted to pnm with pngtopnm (from imagemagick))

resources/fonts/generated_extract_fail/

Images generated incorrectly (font with hyphens). The directory is ignored.

./mano/

Images belonging to the old ocr (OCRv1, see redemption project).

resources/dict_*.trie

Generated from resources/dict_*.txt by make_trie_dictionary

resources/glyphs/*

Generated from image files (font/*) by write_glyphs_bin and normalizer_glyphs_bin

resources/datas/*

Generated from data glyphs (glyphs/*) by write_datas