tags : Machine Learning, Image Compression, Computer Vision

Comparision

TypeNameDescription
ServiceClaude/OpenAI/AWSThey have APIs
LSTM-CNNTesseract
PP-OCR(DB+CRNN)PaddleOCRWorks with rotated stuff
EasyOCR
Toolbox, Modular modelsdoctrSome people mention it works better than paddle and tesseract.
Pytorch+mmlabsMMOCRMight be nice if using mmdetection stuff
suryaOnly for documents, doesn’t work in handwritten. faster than tesseract, Language support. Tries to guess proper reading order.
VLMTrOCR
VLMDONUT
VLMInternVL
VLMIdefics2

Resources