Research map / Detection and segmentation / Open-set and few-shot detection
Rethinking Text Segmentation: A Novel Dataset and a Text-Specific Refinement Approach
Xingqian Xu, Zhifei Zhang, Zhaowen Wang, Brian Price, Zhonghao Wang, Humphrey Shi
Check 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 |
|---|---|---|---|
| Apply TexRNet refinement to HRNetV2-W48 | HRNetV2-W48 · TextSeg | fgIoU: 85.98 -> 86.84 | Improved |
| Apply TexRNet refinement to HRNetV2-W48 | HRNetV2-W48 · ICDAR13 FST | fgIoU: 72.45 -> 73.38 | Improved |
| Apply TexRNet refinement to DeeplabV3+ | DeeplabV3+ · COCO_TS | fgIoU: 72.07 -> 73.98 | Improved |
| Apply TexRNet refinement to DeeplabV3+ | DeeplabV3+ · MLT_S | fgIoU: 84.63 -> 86.31 | Improved |
| Apply TexRNet refinement to HRNetV2-W48 | HRNetV2-W48 · Total-Text | fgIoU: 76.23 -> 78.47 | Improved |
| Add key features pooling and attention (ablation) | TexRNet (base) with DeeplabV3+ · TextSeg | fgIoU: 84.86 -> 85.36 | Improved |
| Add trimap loss (ablation) | TexRNet with attention · TextSeg | fgIoU: 85.36 -> 85.55 | Improved |
| Add glyph discriminator (ablation) | TexRNet with attention and trimap loss · TextSeg | fgIoU: 85.55 -> 86.06 | Improved |
| Train with TextSeg in addition to ICDAR13 FST, COCO_TS, and MLT_S | TexRNet · ICDAR13 FST test | fgIoU: 76.68 -> 78.65 | Improved |
| Train with TextSeg in addition to Total-Text, COCO_TS, and MLT_S | TexRNet · Total-Text test | fgIoU: 77.40 -> 80.01 | Improved |
Closest papers by meaning
- BTS: A Bi-Lingual Benchmark for Text Segmentation in the Wild (CVPR 2022 · similarity 0.89)
- WAS: Dataset and Methods for Artistic Text Segmentation (ECCV 2024 · similarity 0.88)
- A Simple Framework for Text-Supervised Semantic Segmentation (CVPR 2023 · similarity 0.87)
- Image-Text Co-Decomposition for Text-Supervised Semantic Segmentation (CVPR 2024 · similarity 0.86)
- DocSAM: Unified Document Image Segmentation via Query Decomposition and Heterogeneous Mixed Learning (CVPR 2025 · similarity 0.85)
- Improving Long-Text Alignment for Text-to-Image Diffusion Models (ICLR 2025 · similarity 0.84)
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.
