Optical Character Recognition (OCR) on Benchmarking Chinese Text Recognition: Datasets, Baselines, and an Empirical Study

Metric: Accuracy (%) (higher is better)

LeaderboardDataset

Loading chart...

Results

Submit a result

Sort:

#	Model↕	Accuracy (%)▼	Augmentations	Paper	Date↕	Code
1	DTrOCR	89.6	No	DTrOCR: Decoder-only Transformer for Optical Cha...	2023-08-30	Code
2	DTrOCR 105M	89.6	No	DTrOCR: Decoder-only Transformer for Optical Cha...	2023-08-30	Code
3	MaskOCR-L	82.6	No	MaskOCR: Text Recognition with Masked Encoder-De...	2022-06-01	-
4	TransOCR	72.8	No	-	-	Code
5	SRN	65	No	Towards Accurate Scene Text Recognition with Sem...	2020-03-27	Code
6	MORAN	64.3	No	A Multi-Object Rectified Attention Network for S...	2019-01-10	Code
7	SEED	61.2	No	SEED: Semantics Enhanced Encoder-Decoder Framewo...	2020-05-22	Code

#1DTrOCRSOTA
89.6
Accuracy (%)· 2023-08-30
DTrOCR: Decoder-only Transformer for Optical Character Recognition Code
#2DTrOCR 105M
89.6
Accuracy (%)· 2023-08-30
DTrOCR: Decoder-only Transformer for Optical Character Recognition Code
#3MaskOCR-LSOTA
82.6
Accuracy (%)· 2022-06-01
MaskOCR: Text Recognition with Masked Encoder-Decoder Pretraining
#4TransOCR
72.8
Accuracy (%)
No paperCode
#5SRNSOTA
65
Accuracy (%)· 2020-03-27
Towards Accurate Scene Text Recognition with Semantic Reasoning Networks Code
#6MORANSOTA
64.3
Accuracy (%)· 2019-01-10
A Multi-Object Rectified Attention Network for Scene Text Recognition Code
#7SEED
61.2
Accuracy (%)· 2020-05-22
SEED: Semantics Enhanced Encoder-Decoder Framework for Scene Text Recognition Code