Scene Text Recognition2020en

cute80

Dataset from Papers With Code

Metrics:accuracy, cer, wer, f1
Current State of the Art

CPPD

Unknown

99.7

accuracy

accuracy Progress Over Time

Showing 5 breakthroughs from Aug 2021 to May 2023

91.193.595.898.1100.5Aug 2021Jan 2022Jun 2022Nov 2022May 2023accuracyDate

Key Milestones

Aug 2021
DPAN

From paper: Look Back Again: Dual Parallel Attention Network for Accurate and Robust Scene Text Recognition

91.9
Nov 2021
MATRN

From paper: Multi-modal Text Recognition Networks: Interactive Enhancements between Visual and Semantic Features

93.5
+1.7%
Dec 2021
S-GTR

From paper: Visual Semantics Allow for Textual Reasoning Better in Scene Text Recognition

94.7
+1.3%
Sep 2022
MGP-STR

From paper: Multi-Granularity Prediction for Scene Text Recognition

99.3
+4.9%
May 2023
CLIP4STR-L (DataComp-1B)Current SOTA

From paper: CLIP4STR: A Simple Baseline for Scene Text Recognition with Pre-trained Vision-Language Model

99.7
+0.4%
Total Improvement
8.5%
Time Span
1y 9m
Breakthroughs
5
Current SOTA
99.7

Top Models Performance Comparison

Top 10 models ranked by accuracy

accuracy1CPPD99.7100.0%2CLIP4STR-L (DataComp-1B)99.7100.0%3MGP-STR99.399.6%4CLIP4STR-B*99.399.6%5DTrOCR 105M99.199.4%6CLIP4STR-L99.099.3%7CCD-ViT-Small(ARD_2.8M)98.398.6%8CCD-ViT-Base(ARD_2.8M)98.398.6%9CCD-ViT-Tiny(ARD_2.8M)95.896.1%10S-GTR94.795.0%0%25%50%75%100%% of best
Best Score
99.7
Top Model
CPPD
Models Compared
10
Score Range
5.0

accuracyPrimary

#ModelScorePaper / CodeDate
1
CPPD
99.7Jul 2023
2
CLIP4STR-L (DataComp-1B)
99.7May 2023
3
MGP-STR
99.31Sep 2022
4
CLIP4STR-B*
99.3May 2023
5
DTrOCR 105M
99.1Aug 2023
6
CLIP4STR-L
99May 2023
7
CCD-ViT-Small(ARD_2.8M)
98.3Nov 2022
8
CCD-ViT-Base(ARD_2.8M)
98.3Nov 2022
9
CCD-ViT-Tiny(ARD_2.8M)
95.8Nov 2022
10
S-GTR
94.7Dec 2021
11
MATRN
93.5Nov 2021
12
SIGA_T
93.1Mar 2022
13
DiffusionSTR
92.5Jun 2023
14
NRTR+TPS++
92.4May 2023
15
DPAN
91.9
Look Back Again: Dual Parallel Attention Network for Accurate and Robust Scene Text RecognitionCode
Aug 2021
16
CDistNet (Ours)
89.58Nov 2021

Related Papers11

Other Scene Text Recognition Datasets