Research map / Detection and segmentation / Open-set and few-shot detection
Two-Level Transformer and Auxiliary Coherence Modeling for Improved Text Segmentation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 auxiliary coherence objective (ablation) | TLT-TS · Wiki-727K | Pk error: 19.41 -> 15.95 | Helped |
| add auxiliary coherence objective (ablation) | TLT-TS · Wiki-50 | Pk error: 17.47 -> 16.53 | Helped |
| add auxiliary coherence objective (ablation) | TLT-TS · Choi | Pk error: 23.26 -> 18.50 | Helped |
| add auxiliary coherence objective (ablation) | TLT-TS · Cities | Pk error: 19.21 -> 16.85 | Helped |
| add auxiliary coherence objective (ablation) | TLT-TS · Elements | Pk error: 20.33 -> 18.41 | Helped |
| add auxiliary coherence objective (ablation) | TLT-TS · Wiki-50-CS zero-shot | Pk error: 24.27 -> 22.32 | Helped |
| add auxiliary coherence objective (ablation) | TLT-TS · Wiki-50-FI zero-shot | Pk error: 25.99 -> 22.87 | Helped |
| add auxiliary coherence objective (ablation) | TLT-TS · Wiki-50-TR zero-shot | Pk error: 25.89 -> 24.20 | Helped |
| replace BiLSTM hierarchy with two-level Transformer | TLT-TS · Elements | Pk error: 41.63 -> 20.33 | Helped |
| supervised CATS vs unsupervised GraphSeg | CATS · Choi | Pk error: 18.50 vs 5.6--7.2 | Hurt |
Closest papers by meaning
- Image-Text Co-Decomposition for Text-Supervised Semantic Segmentation (CVPR 2024 · similarity 0.86)
- BBScore: A Brownian Bridge Based Metric for Assessing Text Coherence (AAAI 2024 · similarity 0.85)
- Hierarchical Coherence Modeling for Document Quality Assessment (AAAI 2021 · similarity 0.85)
- CAT-Seg: Cost Aggregation for Open-Vocabulary Semantic Segmentation (CVPR 2024 · similarity 0.84)
- Text4Seg: Reimagining Image Segmentation as Text Generation (ICLR 2025 · similarity 0.83)
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.83)
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.