nguyenq / tess4j

Java JNA wrapper for Tesseract OCR API
Apache License 2.0
1.61k stars 373 forks source link

The recognition results of tess4j and tesseract are inconsistent. #176

Closed shihuili1218 closed 4 years ago

shihuili1218 commented 4 years ago

tess4j version: 4.5.1 tesseract version: 4.1.1 leptonica version: 1.79 system: centos 7

I saw the Change Log on the tess4j website[http://tess4j.sourceforge.net/changelog.html], where tess4j 4.5.1 USES Leptonica 1.79.0 and Tesseract 4.1.1. But is it normal that I get inconsistent results using tess4j and Leptonica + Tesseract?

test data: http://q3820ps1a.bkt.clouddn.com/test/S00105-15295112.png traineddata: https://github.com/tesseract-ocr/tessdata/blob/master/chi_sim.traineddata

nguyenq commented 4 years ago

If you call the same API methods with the same parameters, you'd get the same results.

nguyenq commented 4 years ago

Can the ticket be closed?

shihuili1218 commented 4 years ago

Can the ticket be closed?

http://qiniu.ofcoder.com/test/S00105-15295112.png