AI Paper Atlas

Research map / Benchmarks / SVT

SVT leaderboard

31 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on SVT. Best accuracy: 97.8 (DARLING, CVPR 2024).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

SVT · accuracy

21 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
197.8DARLINGChoose What You Need: Disentangled Representation Learning for Scene Text Recognition Removal and Editing
CVPR 2024 · Table 1
Editing and style transfer
296.3TRBA-cr
TRBA
Pushing the Performance Limit of Scene Text Recognizer Without Human Annotation
CVPR 2022 · Table 1
Semi-supervised learning
394.6CornerTransformerToward Understanding WordArt: Corner-Guided Transformer for Scene Text Recognition
ECCV 2022 · Table 7
Human and embodied modeling
494.3ABINet-ConCLR
ABINet
Context-Based Contrastive Learning for Scene Text Recognition
AAAI 2022 · Table 3
Human and embodied modeling
593.5ABINet-LVRead Like Humans: Autonomous, Bidirectional and Iterative Language Modeling for Scene Text Recognition
CVPR 2021 · Table: Accuracy comparison wit
Human and embodied modeling
693.2GTC (1D) + GCN
ResNet50
GTC: Guided Training of CTC towards Efficient and Accurate Scene Text Recognition
AAAI 2020 · Table 2
CNN features
792.7SCATTER (5 Block)SCATTER: Selective Context Attentional Scene Text Recognizer
CVPR 2020 · Table 1
Attention and transformers
891.7VisionLANFrom Two to One: A New Scene Text Recognizer With Visual Language Modeling Network
ICCV 2021 · Table 5
Human and embodied modeling
991.5SRN
ResNet50+FPN
Towards Accurate Scene Text Recognition With Semantic Reasoning Networks
CVPR 2020 · Table: Comparisons of scene te
Human and embodied modeling
1091.3TRBA-PR
TRBA
What if We Only Use Real Datasets for Scene Text Recognition? Toward Scene Text Recognition With Fewer Labels
CVPR 2021 · Table: Accuracy of STR models
Semi-supervised learning
11 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

SVT · word accuracy (%)

10 papers, each with its best reported number for its own method; higher is better.

#word accuracy (%) ↑MethodPaperTopic
198.76CLIP4STR-B*
CLIP4STR-B
An Empirical Study of Scaling Law for Scene Text Recognition
CVPR 2024 · Table 11
RNNs and normalization
298.61E2STR-ICL
ViT + OPT-125M
Multi-modal In-Context Learning Makes an Ego-evolving Scene Text Recognizer
CVPR 2024 · Table 1
Human and embodied modeling
397.7OursTowards Ship License Plate Recognition in the Wild: A Large Benchmark and Strong Baseline
AAAI 2025 · Table 6
Self-supervised and contrastive
495.8S-GTR
Segmentation-based VR + SRN LM
Visual Semantics Allow for Textual Reasoning Better in Scene Text Recognition
AAAI 2022 · Table 1
Human and embodied modeling
595MATRNMulti-modal Text Recognition Networks: Interactive Enhancements between Visual and Semantic Features
ECCV 2022 · Table 1
Cross-modal learning
694PREN2DPrimitive Representation Learning for Scene Text Recognition
CVPR 2021 · Table 1
Human and embodied modeling
790.9AutoSTRAutoSTR: Efficient Backbone Search for Scene Text Recognition
ECCV 2020 · Table 2
Neural architecture search
886.8PerSec-ViT + UTI-100M
ViT
Perceiving Stroke-Semantic Context: Hierarchical Contrastive Learning for Robust Scene Text Recognition
AAAI 2022 · Table 3
Self-supervised and contrastive
986.6Recognizer on generated images (full method)
ResNet-45 attention recognizer
Separating Content from Style Using Adversarial Learning for Recognizing Text in the Wild
IJCV 2021 · Table 1
GANs
1058.6Ours (blurred)SimAN: Exploring Self-Supervised Representation Learning of Scene Text via Similarity-Aware Normalization
CVPR 2022 · Table: Word accuracy on benchm
Editing and style transfer

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.