If you think those images are visually identical you should visit your optician. :-)
The ImageMagick version is much blurrier, so I'd guess that the high frequency noise from the pixelated OpenCV image is making Tesseract unhappy. If you want to continue using OpenCV, try applying a Gaussian Blur after whatever other operations you're using and see if that creates an image which matches the ImageMagick one. Tom On Sunday, October 4, 2015 at 11:11:54 PM UTC-4, Bhaarat Sharma wrote: > > I have two versions of the same image. One version was resized and rotated > using OpenCV, however, the other was resized and rotated using ImageMagick. > > Tesseract finds text from the ImageMagick version of the image but not > from OpenCV version of the image. I don't know why this is. > > Here are the details of the two images using the identify command from > imagemagick: > > $ identify original_zoomed_rotated_im.png > original_zoomed_rotated_im.png PNG 1334x776 1334x776-34-76 8-bit sRGB > 256c 131KB 0.000u 0:00.000 > $ identify opencv_zoomed_rotated.png > opencv_zoomed_rotated.png PNG 1334x776 1334x776-34-76 8-bit sRGB 256c > 47.4KB 0.000u 0:00.000 > > I've attached the two images > > Here are the outputs from tesseract > > $ tesseract opencv_zoomed_rotated.png ozr && more ozr.txt > Tesseract Open Source OCR Engine v3.02.02 with Leptonica > $ tesseract original_zoomed_rotated_im.png ozr_im && more ozr_im.txt > Tesseract Open Source OCR Engine v3.02.02 with Leptonica > MORTE 5;‘ > l‘ PARA > ROUSSEFF > > > > -- 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 http://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/d0b229b1-af3c-4bb8-b1bd-85327ffb3ff1%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
