Home Console Docs News Papers Tutorials Datasets Wiki SOTA LLM Models GPU Leaderboard Events

English

Scene Text Detection On Icdar 2013

Metrics

F-Measure

Precision

Recall

Results

Performance results of various models on this benchmark

				Paper Title	Repository
TextFuseNet (ResNeXt-101)	94.61%	97.27	92.09	TextFuseNet: Scene Text Detection with Richer Fused Features	-
SPCNET	92.1%	93.8	90.5	Scene Text Detection with Supervised Pyramid Context Network
Mask TextSpotter	91.7%	95	88.6	Mask TextSpotter: An End-to-End Trainable Neural Network for Spotting Text with Arbitrary Shapes
WordSup (VGG16-synth-icdar)	90.34%	93.34	87.53	WordSup: Exploiting Word Annotations for Character based Text Detection	-
STN-OCR	90.3%	-	-	STN-OCR: A single Neural Network for Text Detection and Text Recognition
PixelLink+VGG16 2s MS	88.1%	88.6	87.5	PixelLink: Detecting Scene Text via Instance Segmentation
Corner Localization (multi-scale)	88%	92	84.4	Multi-Oriented Scene Text Detection via Corner Localization and Region Segmentation
TextBoxes++_MS	88%%	91	84	TextBoxes++: A Single-Shot Oriented Scene Text Detector
Corner-based Region Proposals	87.6%%	91.9	83.9	Detecting Multi-Oriented Text with Corner-based Region Proposals
SSTD	87%	88	86	Single Shot Text Detector with Regional Attention
SegLink	85.3%	87.7	83	Detecting Oriented Text in Natural Images by Linking Segments
Gupta et al.	83.0%	92.0	75.5	Synthetic Data for Text Localisation in Natural Images
USM (COCO TS + ICDAR–2013)	80.40%	-	-	Unsharp Masking Layer: Injecting Prior Knowledge in Convolutional Networks for Image Classification	-
Neumann et al. *	77.1%	81.8	72.4	Efficient Scene Text Localization and Recognition with Local Character Refinement	-
Jaderberg et al.	76.8%	88.5	67.8	Reading Text in the Wild with Convolutional Neural Networks	-
CRAFT	-	97.4	93.1	Character Region Awareness for Text Detection

0 of 16 row(s) selected.