Lorenzo

If this can be of any help to someone, I got the exact same error when I run the clstmocrtrain without the corresponding *.gt.txt files in the image folder.

The confidence value is not a "normal" percentage. 93 does not mean that there is the 93% of probabilities for that being a face (93% of what?). With many different...

Ciao @rawmain e grazie per la risposta. La mia domanda allora si puo' riformulare in questo modo: e' possibile vedere gia' da ora quali saranno le validation rules dal 6...

Hi sirfz, thanks for you answer. I'm building leptonica right now. Just one question, does the example works for you?

In the meantime I found a clue. I rebuilt everything, and I have the same problem, but I noticed that if I call: api.Init(lang="eng") # or any other language before...

Thanks. I'll try with 4.0 too and report back. You also get a much higher score than I do with my workaround (74). And the fact that you do not...

Can I ask you one more favor (even if this is quite likely a tesseract problem)? I added these two lines to my script: th = api.GetThresholdedImage() th.save('th_tess_bind.png') just before...

@nttstar Thanks, resizing and padding solved the problem for me. Or at least made it much more manageable. I resize the largest dimension to the closest one from a fixed...

@woreom Just pad without resizing: pad_sizes = np.asarray([150, 250, 300, 350, 500, 600, 720, ..., 4096, 5000, 6000, 7000, 8000, 9000, 10000]) def pad_image(img): h, w, *_ = img.shape if...

Replace the three dots in the pad_sizes array with the values you want to use: 800, 1000, etc. The list was too long so I removed some values and put...