#Want: OCR app on Phone for Converting Books to Text (possibly for text-to-speech), but Continually Re-train and Re-Run the OCR Algorithm via Bayesian Stats => OCR that might start out poor, but gets better, especially if you scan more pages from same text