Scene Text Detection On Total Text

평가 지표

F-Measure

FPS

Precision

Recall

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	F-Measure	FPS	Precision	Recall	Paper Title	Repository
DPText-DETR (ResNet-50)	89.0%	17	91.8%	86.4%	DPText-DETR: Towards Better Scene Text Detection with Dynamic Points in Transformer	-
FAST-T-448	81.6%	152.8	86.5	77.2	FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation	-
Ch,ng et al.	36.0%	-	40.0	33.0	Total-Text: A Comprehensive Dataset for Scene Text Detection and Recognition	-
PSENet-4s	79.6%	-	84.5	75.2	Shape Robust Text Detection with Progressive Scale Expansion Network	-
CharNet H-88	85.6%	-	89.9	81.7	Convolutional Character Networks	-
CRAFT	83.6%	-	87.6	79.9	Character Region Awareness for Text Detection	-
Mask TextSpotter	61.3%	-	69	55	Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes	-
EAST	42.0%	-	50.0	36.2	EAST: An Efficient and Accurate Scene Text Detector	-
DB-ResNet-50 (800)	84.7%	-	-	-	Real-time Scene Text Detection with Differentiable Binarization	-
TextFuseNet (ResNeXt-101)	87.5%	-	89.2	85.8	TextFuseNet: Scene Text Detection with Richer Fused Features
CharNet H-88 (multi-scale)	86.5%	-	88	85	Convolutional Character Networks	-
FAST-B-512	85.8%	93.2	89.6	82.4	FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation	-
FAST-S-512	84.9%	115.5	88.3	81.7	FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation	-
SRFormer (ResNet-50)	90.0%	-	92.2%	87.9%	SRFormer: Text Detection Transformer with Incorporated Segmentation and Regression	-
FTSN	81.3%	-	84.7	78	Fused Text Segmentation Networks for Multi-oriented Scene Text Detection	-
TextFiled	80.6%	-	81.2	79.9	TextField: Learning A Deep Direction Field for Irregular Scene Text Detection	-
TextSnake	78.4%	-	82.7	74.5	TextSnake: A Flexible Representation for Detecting Text of Arbitrary Shapes	-
TextCohesion	84.6%	-	-	-	TextCohesion: Detecting Text for Arbitrary Shapes	-
MixNet	90.5%	15.2	93.0	88.1	MixNet: Toward Accurate Detection of Challenging Scene Text in the Wild	-
SPCNET	82.9%	-	83	82.8	Scene Text Detection with Supervised Pyramid Context Network	-

0 of 27 row(s) selected.