Research map / Detection and segmentation / Open-set and few-shot detection
Advancing Referring Expression Segmentation Beyond Single Image
Yixuan Wu, Zhao Zhang, Chi Xie, Feng Zhu, Rui Zhao
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 Triphasic Query Module (intra-group visual queries) (ablation) | GRSer · G-RefCg (GRES) | adapted mIoU: 66.28 -> 75.32 | Improved |
| add Heatmap Hierarchizer ranking (ablation) | GRSer · G-RefCg (GRES) | adapted mIoU: 68.92 -> 75.32 | Improved |
| add mirror training strategy (ablation) | GRSer · G-RefCg (GRES) | adapted mIoU: 69.38 -> 75.32 | Improved |
| add triplet margin loss (ablation) | GRSer · G-RefCg (GRES) | adapted mIoU: 30.37 -> 75.32 | Improved |
| increase group size from 1 to 8 (ablation) | GRSer TQM · G-RefCg (GRES) | adapted mIoU: 66.28 -> 75.32 | Improved |
| use consistent combined ranking Rpos+Rneg in train and test instead of random (ablation) | Heatmap Hierarchizer · G-RefCg (GRES) | adapted mIoU: 68.92 -> 75.32 | Improved |
| combine positive and negative ranks instead of one (ablation) | Heatmap Hierarchizer · G-RefCg (GRES) | adapted mIoU: 74.57 -> 75.32 | Improved |
| group-wise processing vs single-image RES methods | GRSer vs LAVT · GRD (GRES) | adapted mIoU: 18.48 -> 57.12 | Improved |
| GRSer on RES setting without negative samples | GRSer vs LAVT · G-RefCg (RES) | mIoU: 59.94 -> 65.47 | Improved |
| GRSer on CoCA E-measure vs GCoNet | GRSer · CoCA | E-measure: 0.739 -> 0.728 | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- DeepACG — DeepACG: Co-Saliency Detection via Semantic-Aware Contrast Gromov-Wasserstein Distance (CVPR 2021) · better on 3 of 3 shared results
- GCoNet — GCoNet+: A Stronger Group Collaborative Co-Salient Object Detector (TPAMI 2023) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- GRES: Generalized Referring Expression Segmentation (CVPR 2023 · similarity 0.95)
- Unveiling Parts Beyond Objects: Towards Finer-Granularity Referring Expression Segmentation (CVPR 2024 · similarity 0.91)
- GSVA: Generalized Segmentation via Multimodal Large Language Models (CVPR 2024 · similarity 0.91)
- RESAnything: Attribute Prompting for Arbitrary Referring Segmentation (NeurIPS 2025 · similarity 0.89)
- Mask Grounding for Referring Image Segmentation (CVPR 2024 · similarity 0.89)
- RefDetector: A Simple Yet Effective Matching-based Method for Referring Expression Comprehension (AAAI 2025 · similarity 0.88)
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.
