Optical Character Recognition On Benchmarking

Accuracy (%)

Results

Performance results of various models on this benchmark

Model Name	Accuracy (%)	Paper Title	Repository
DTrOCR 105M	89.6	DTrOCR: Decoder-only Transformer for Optical Character Recognition
TransOCR	72.8	Scene Text Telescope: Text-Focused Scene Image Super-Resolution
MaskOCR-L	82.6	MaskOCR: Text Recognition with Masked Encoder-Decoder Pretraining	-
DTrOCR	89.6	DTrOCR: Decoder-only Transformer for Optical Character Recognition
SEED	61.2	SEED: Semantics Enhanced Encoder-Decoder Framework for Scene Text Recognition
SRN	65.0	Towards Accurate Scene Text Recognition with Semantic Reasoning Networks
MORAN	64.3	A Multi-Object Rectified Attention Network for Scene Text Recognition

0 of 7 row(s) selected.