Research map / Detection and segmentation / Weakly supervised detection and segmentation
Learning To Segment Every Referring Object Point by Point
Mengxue Qu, Yu Wu, Yunchao Wei, Wu Liu, Xiaodan Liang, Yao 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 |
|---|---|---|---|
| Use the proposed method instead of the baseline with 1% mask annotations | RES sequence prediction model · RefCOCO+ testA | IoU: 48.81 -> 52.06 | Improved |
| Use the proposed method instead of the baseline with 5% mask annotations | RES sequence prediction model · RefCOCO+ testA | IoU: 54.22 -> 54.68 | Improved |
| Use the proposed method instead of the baseline with 10% mask annotations | RES sequence prediction model · RefCOCO+ testA | IoU: 54.87 -> 55.79 | Improved |
| Add co-content teacher-forcing (CCTF) (ablation) | SeqTR-based RES model with 1% mask labels · RefCOCOg val | IoU: 42.57 -> 44.20 (+1.63) | Improved |
| Add point-modulated cross-attention (PMCA) (ablation) | SeqTR-based RES model with 1% mask labels and CCTF · RefCOCOg val | IoU: 44.20 -> 44.58 (+0.38) | Improved |
| Add co-content teacher-forcing (CCTF) (ablation) | SeqTR-based RES model with 100% mask labels · RefCOCOg val | IoU: 55.95 -> 56.03 (+0.08) | No clear effect |
| Add point-modulated cross-attention (PMCA) (ablation) | SeqTR-based RES model with 100% mask labels and CCTF · RefCOCOg val | IoU: 56.03 -> 56.89 (+0.86) | Improved |
| Resample predicted pseudo points uniformly rather than use predicted points directly (ablation) | CCTF + PMCA RES model during pseudo-label retraining · RefCOCOg val | IoU: 44.27 -> 46.80 | Improved |
| Select pseudo masks using threshold T=0.10 instead of using all predictions (ablation) | RPP pseudo-label retraining · RefCOCOg val | IoU: 44.53 -> 46.80 | Improved |
| Set pseudo-mask selection threshold T to 0.05 rather than T=0.10 (ablation) | RPP pseudo-label retraining · RefCOCOg val | IoU: 46.80 -> 44.35 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- ResTR — ReSTR: Convolution-Free Referring Image Segmentation Using Transformers (CVPR 2022) · 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
- Unveiling Parts Beyond Objects: Towards Finer-Granularity Referring Expression Segmentation (CVPR 2024 · similarity 0.92)
- Curriculum Point Prompting for Weakly-Supervised Referring Image Segmentation (CVPR 2024 · similarity 0.91)
- Mask Grounding for Referring Image Segmentation (CVPR 2024 · similarity 0.90)
- SAM as the Guide: Mastering Pseudo-Label Refinement in Semi-Supervised Referring Expression Segmentation (ICML 2024 · similarity 0.90)
- OneRef: Unified One-tower Expression Grounding and Segmentation with Mask Referring Modeling (NeurIPS 2024 · similarity 0.89)
- Rethinking Two-Stage Referring Expression Comprehension: A Novel Grounding and Segmentation Method Modulated by Point (AAAI 2024 · similarity 0.89)
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.
