Research map / Vision-language and multimodal / Human and embodied modeling
Multi-Granularity Prediction with Learnable Fusion for Scene Text Recognition
Cheng Da, Peng Wang, Cong Yao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Character A3 module (ablation) | MGP-STR ViT · 6 STR benchmarks AVG | accuracy: 91.58 -> 92.73 | Improved |
| use 32x128 input with 4x4 patches instead of 224x224 with 16x16 (ablation) | MGP-STR ViT · 6 STR benchmarks AVG | accuracy: 91.07 -> 91.58 | Improved |
| add BPE and WordPiece heads with CFS fusion (ablation) | MGP-STR Vision · 6 STR benchmarks AVG | accuracy: 92.73 -> 93.35 | Improved |
| use Cumprod instead of Mean confidence score (ablation) | MGP-STR CFS · 6 STR benchmarks AVG | accuracy: 93.28 -> 93.35 | Improved |
| add BPE head only (C+B) with CFS (ablation) | MGP-STR Vision · 6 STR benchmarks AVG | accuracy: 92.73 -> 93.24 | Improved |
| add WordPiece head only (C+W) with CFS (ablation) | MGP-STR Vision · 6 STR benchmarks AVG | accuracy: 92.73 -> 93.25 | Improved |
| replace explicit BCN language model with multi-granularity prediction (ablation) | MGP-STR Vision · 6 STR benchmarks AVG | accuracy: 93.14 -> 93.35 | Improved |
| fuse heads with CFS on smaller DeiT-Tiny backbone (ablation) | MGP-STR · 6 STR benchmarks AVG | accuracy: 89.36 -> 89.91 | Improved |
| fuse heads with CFS on DeiT-Small backbone (ablation) | MGP-STR · 6 STR benchmarks AVG | accuracy: 91.01 -> 91.96 | Improved |
| swap DeiT init for MAE, DINO, DINOv2, CLIP or BLIP (ablation) | MGP-STR CFS (200M iters) · 6 STR benchmarks AVG | accuracy: 93.70 vs 93.60, 93.47, 93.59, 93.60, 93.53 | No clear effect |
| replace CFS with Learnable Fusion Strategy (ablation) | MGP-STR (200M iters) · 6 STR benchmarks AVG | accuracy: 93.70 -> 94.05 | Improved |
| increase LFS noised text from 8 to 256 (ablation) | MGP-STR LFS · 6 STR benchmarks AVG | accuracy: 93.67 -> 94.05 | Improved |
Closest papers by meaning
- Multi-Granularity Prediction for Scene Text Recognition (ECCV 2022 · similarity 0.98)
- Revisiting Scene Text Recognition: A Data Perspective (ICCV 2023 · similarity 0.92)
- What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis (ICCV 2019 · similarity 0.90)
- OTE: Exploring Accurate Scene Text Recognition Using One Token (CVPR 2024 · similarity 0.88)
- Boosting Semi-Supervised Scene Text Recognition via Viewing and Summarizing (NeurIPS 2024 · similarity 0.87)
- Linguistics-aware Masked Image Modeling for Self-supervised Scene Text Recognition (CVPR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
