Research map / Vision-language and multimodal / Human and embodied modeling
Linguistic Structure Guided Context Modeling for Referring Image Segmentation
Tianrui Hui, Si Liu, Shaofei Huang, Guanbin Li, Sansi Yu, Faxi Zhang, Jizhong Han
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 LSCM module (ablation) | RRN-CNN single-level · UNC val | Overall IoU (%): 47.26 -> 54.87 | Improved |
| add LSCM module (ablation) | Multi-Level-RRN-CNN · UNC val | Overall IoU (%): 57.61 -> 59.67 | Improved |
| add GloVe embeddings (ablation) | LSCM single-level · UNC val | Overall IoU (%): 54.87 -> 55.93 | Improved |
| add Mutan fusion (ablation) | LSCM + GloVe single-level · UNC val | Overall IoU (%): 55.93 -> 56.50 | Improved |
| add GloVe embeddings (ablation) | LSCM multi-level · UNC val | Overall IoU (%): 59.67 -> 60.98 | Improved |
| add Dual-Path Fusion instead of Concat Fusion (ablation) | LSCM multi-level · UNC val | Overall IoU (%): 60.10 -> 61.54 | Improved |
| add Dual-Path Fusion instead of Gated Fusion (ablation) | LSCM multi-level · UNC val | Overall IoU (%): 60.46 -> 61.54 | Improved |
| use LSCM instead of GloRe (ablation) | multi-level baseline · UNC val | Overall IoU (%): 58.53 -> 59.67 | Improved |
| use 1 graph conv layer instead of 0 (ablation) | LSCM + GloVe · UNC val | Overall IoU (%): 54.51 -> 55.93 | Improved |
| increase graph conv layers from 1 to 2 (ablation) | LSCM + GloVe · UNC val | Overall IoU (%): 55.93 -> 50.77 | Hurt |
| use adaptive layers equal to DPT depth instead of 1 (ablation) | LSCM + GloVe · UNC val | Overall IoU (%): 55.93 -> 54.69 | Hurt |
| increase graph conv layers from 1 to 2 (ablation) | LSCM + GloVe · G-Ref val | Overall IoU (%): 40.54 -> 39.29 | Hurt |
Closest papers by meaning
- Referring Image Segmentation via Cross-Modal Progressive Comprehension (CVPR 2020 · similarity 0.90)
- Recurrent Multimodal Interaction for Referring Image Segmentation (ICCV 2017 · similarity 0.90)
- Cross-Modal Progressive Comprehension for Referring Segmentation (TPAMI 2021 · similarity 0.89)
- Mask Grounding for Referring Image Segmentation (CVPR 2024 · similarity 0.88)
- LQMFormer: Language-aware Query Mask Transformer for Referring Image Segmentation (CVPR 2024 · similarity 0.87)
- IteRPrimE: Zero-shot Referring Image Segmentation with Iterative Grad-CAM Refinement and Primary Word Emphasis (AAAI 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.
