Currently building an OCR pipeline using CNN + RNN + CTC Loss, inspired by EasyOCR, to extract text from images efficiently.