[tesseract-ocr] Training Tesseract, results unexpected

Hi, I'm trying to use Tesseract to recognise some characters from medical images, and I used JtessBoxEditor to edit box file and adapt it to errors in this image. But when I run tesseract after modify that box whit correct chars, it keeps making an output with previous box values. I don't know

[tesseract-ocr] I can't train tesseract with this medical image

I get this log when I train it specifying the missed squares with JTessBoxEditor: [C:\Users\FlcUser\Desktop\JTess\jTessBoxEditor\tesseract-ocr/tesseract, image_part_003.png, image_part_003, box.train] Tesseract Open Source OCR Engine v3.05.01 with Leptonica FAIL! APPLY_BOXES: boxfile line 1/7 (