I use wrapper tess4j for tesseract 3 to make OCR on cyrillic characters. Before I execute image pre-processing: skewed image, remove background and noise, adjust image quality and finaly get ROI for OCR. Image pre-processing takes 0.5-1 seconds. But tesseract OCR process takes a long time.
How can I enhance tesseract OCR speed perfomance?