Character Region Awareness for Text Detection

Scene text detection methods based on neural networks have emerged recently and have shown promising results. Previous methods trained with rigid word-level bounding boxes exhibit limitations in representing the text region in an arbitrary shape... (read more)

PDF Abstract CVPR 2019 PDF CVPR 2019 Abstract

Results from the Paper


 Ranked #1 on Scene Text Detection on ICDAR 2013 (Precision metric)

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Scene Text Detection ICDAR 2013 CRAFT Precision 97.4 # 1
Recall 93.1 # 1
H-Mean 95.2 # 1
Scene Text Detection ICDAR 2015 CRAFT F-Measure 86.9 # 16
Precision 89.8 # 15
Recall 84.3 # 16
Scene Text Detection ICDAR 2017 MLT CRAFT Precision 80.6 # 7
Recall 68.2 # 11
F-Measure 73.9% # 7
Scene Text Detection MSRA-TD500 CRAFT Recall 78.2 # 3
Precision 88.2 # 2
H-Mean 82.9 # 1
Scene Text Detection SCUT-CTW1500 CRAFT Precision 86 # 4
Recall 81.1 # 5
H-Mean 83.5 # 1
Scene Text Detection Total-Text CRAFT F-Measure 83.6% # 8
Precision 87.6 # 5
Recall 79.9 # 6

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet