Trying to make recognition with custom trained digits.traineddata - got original files by Shreeshrii, modified makedata.sh with more script fonts to catch specific digits look. Tesseract binaries are 4.0.0 (from UB-Mannheim said as 4.1.0)
Results are quite good, but see some strange behaviour: 1. Quite often recognized data contains more text than really exist. For example, scanned image has 4 digits, while at output I see 5 or even 6 digits. Looks like some text characters got recognized twice. 2. Sometimes absolute clear image area produces junk recognition output - 20-30 digits, where some characters are repeated many times (377777775555, etc) Any ideas? -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/d6f38063-0fe0-42c0-9643-df27878a2097%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
