HyperAI超神経

プラットフォーム

Text Spotting On Total Text

評価指標

F-measure (%) - Full Lexicon

F-measure (%) - No Lexicon

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

			Paper Title
DeepSolo (ViTAEv2-S, TextOCR)	89.6	83.6	DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting
DeepSolo (ResNet-50, TextOCR)	88.7	82.5	DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting
DeepSolo (ResNet-50)	87.0	79.7	DeepSolo: Let Transformer Decoder with Explicit Points Solo for Text Spotting
UNITS	86.0	78.7	Towards Unified Scene Text Spotting based on Sequence Generation
A3S	85.1	79.4	A3S: Adversarial learning of semantic representations for Scene-Text Spotting
SwinTextSpotter	84.1	74.3	SwinTextSpotter: Scene Text Spotting via Better Synergy between Text Detection and Text Recognition
TESTR	83.9	73.3	Text Spotting Transformers
MANGO	83.6	72.9	MANGO: A Mask Attention Guided One-Stage Scene Text Spotter
DEER	83.3	74.8	DEER: Detection-agnostic End-to-End Recognizer for Scene Text Spotting
GLASS	83.0	76.6	GLASS: Global to Local Attention for Scene-Text Spotting
MaskTextSpotter v3	78.4	71.2	Mask TextSpotter v3: Segmentation Proposal Network for Robust Scene Text Spotting
ABCNet v2	78.1	70.4	ABCNet v2: Adaptive Bezier-Curve Network for Real-time End-to-end Text Spotting

0 of 12 row(s) selected.