Hi,
I'm using tesseract 4.00alpha tesseract with leptonica 1.74.1 on Ubuntu 14, I use it to create LSTM files, when I want to make a detection for any image I use both trained data with the LSTM file, and the command is tesseract image.tif
output -l ara --oem 1
the results is good except for one character, so my question is how to enhance the accuracy for a specific character
thanks