Scene Text Detection On Icdar 2015

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
DBNet++ (ResNet-18) (736)	83.1	44	90.1	77.2	Real-Time Scene Text Detection with Differentiable Binarization and Adaptive Scale Fusion
PAN	82.9	-	84	81.9	EAST: An Efficient and Accurate Scene Text Detector
SSTD	77	-	80	73	WordSup: Exploiting Word Annotations for Character based Text Detection	-
CharNet H-88 (multi-scale)	91.55	-	92.65	90.47	Convolutional Character Networks
FOTS	87.99	-	91	85.17	FOTS: Fast Oriented Text Spotting with a Unified Network
TextFuseNet (ResNeXt-101)	92.23	-	93.96	90.56	TextFuseNet: Scene Text Detection with Richer Fused Features
CharNet H-57 (single-scale)	89.66	-	88.88	90.45	Convolutional Character Networks
EAST + PVANET2x RBOX (single-scale)	78.2	-	83.6	73.5	EAST: An Efficient and Accurate Scene Text Detector
Quad_MS	82.9	-	87.8	78.5	TextBoxes++: A Single-Shot Oriented Scene Text Detector
FAST-B-1280	87.1	15.7	89.7	84.6	FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
EAST + PVANET2x RBOX (multi-scale)	80.7	-	83.3	78.3	Single Shot Text Detector with Regional Attention
FAST-B-736	84.7	42.7	88.0	81.7	FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
CharNet H-57 (multi-scale)	90.06	-	91.43	88.74	Convolutional Character Networks
FAST-T-736	81.7	60.9	86	77.9	FAST: Faster Arbitrarily-Shaped Text Detector with Minimalist Kernel Representation
PGNet-E	-	-	-	-	-	-
FOTS MS	89.84	-	91.85	87.92	FOTS: Fast Oriented Text Spotting with a Unified Network
SAST	86.91	-	86.72	87.09	A Single-Shot Arbitrarily-Shaped Text Detector based on Context Attended Multi-Task Learning
PSENet-1s	87.1	-	88.7	85.5	Shape Robust Text Detection with Progressive Scale Expansion Network
DB-ResNet-50 (1152)	87.3	-	91.8	83.2	Real-time Scene Text Detection with Differentiable Binarization
PGNet-A	-	-	-	-	PGNet: Real-time Arbitrarily-Shaped Text Spotting with Point Gathering Network

0 of 43 row(s) selected.