Because I benchmarked both on my dataset and found that Tesseract was better for my use-case?