ALOCR Ver. 1.0, Japanese printed character recognition library is released in 1999. It's based on up-to-date algorithms. We provide a paragraph recognition API, a text line recognition API and dictionary maintenance functions.
You can use ALPRIM Ver. 1.0 library to abstract paragraph or lines from full document images. It can handle complicated layout. It can also abstract text lines from table or graphics. ALPRIM has a function to infer the angle of inclined text lines.

OCR system


  1. High recognition rate
    Development of this library started in 1998. It used up-to-date character recognition algorithms supposed to be used on 21st century computer.
  2. Easy to customize
    You can add pattern dictionary for special purpose.
  3. Powerful language processing
    This library is powerd by a Japanese dictionary over 20 MBytes. The dictionary include statistical probability of Japanese character connectivity.

Character Code

It depends on pattern dictionary. The default dictionary has all kanjis belong to JIS Level 1, major symbol, alpabet, number, katakana, hiragana, high frequent kanjis and "Name Kanji" belong to JIS Level 2.
You can extend pattern dictionary easily by using pattern registration function or pattern dictionary compiler.

Font Type

System dictionary includes two major Japanese font "Mincho" and "Gothic", in addition system dictionary contains "Maru-Gochic" and "Japanese Text-Book" fonts. We provide some optional dictionary. "Kaisho" and "Reisho" pattern dictionary and so on. "Pattern dictionary compiler" can make pattern dictionary from any Japanese TrueType font.

Font Size

On 400 dpi image, ALOCR can recognize over 6-point font size. Font size depends on resolution of aquired image.
Users manual

Reference manual