Multi-lingual mathematical document recognition by InftyReader


M. Suzuki
Japan
Play (29min) Download:  MP4 | MP3

InftyReader is the software developed in Kyushu University to recognize mathematical documents including various formulas of pure and applied mathematics.

It uses commercial OCR engines to recognize ordinary text parts. One of the crucial points to keep high accuracy of the recognition is the segmentation of the text area and math expression area to combine commercial OCR and InftyOCR. Recently, we are trying to use the OCR engine of ABBY FineReader to adapt InftyReader to various European languages.

In the talk, I will briefly sketch the methods to combine different OCR engines and will give some demonstrations of the current state of our New InftyReader.

You may also like:

  1. Mathematical Formulae Recognition
  2. Mathematical Document Classification via Symbol Frequency Analysis
  3. UMCL: Providing Braille Transcription for Mathematical Applications
  4. Drawing SVG in the Cloud
  5. Challenges in implementing a multi-lingual JATS publishing workflow

  • Share
1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading ... Loading ...