Research map / Detection and segmentation / Weakly supervised detection and segmentation
Entity-enhanced Adaptive Reconstruction Network for Weakly Supervised Referring Expression Grounding
Xuejing Liu, Liang Li, Shuhui Wang, Zheng-Jun Zha, Zechao Li, Qi Chuan Tian, Qingming Huang
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 entity enhancement and soft context pooling to ARN (EARN vs ARN) | ARN · RefCOCOg val | accuracy (IoU>0.5): + 10.67% | Improved |
| add entity enhancement and adaptive reconstruction vs IGN+CCL | weak-supervised REG · RefCOCO testB | accuracy (IoU>0.5): 32.59 -> 38.59 | Improved |
| add entity enhancement (selects candidate proposals) (ablation) | EARN on RefCLEF · RefCLEF val | accuracy (IoU>0.5): 26.19 -> 36.68 | Improved |
| add adaptive reconstruction on top of language reconstruction (ablation) | EARN on RefCLEF · RefCLEF val | accuracy (IoU>0.5): 21.86 -> 25.35 | Improved |
| add attribute classification (ablation) | EARN on RefCLEF · RefCLEF val | accuracy (IoU>0.5): 25.35 -> 26.19 | Improved |
| add adaptive reconstruction to entity-enhanced model (ablation) | EARN · RefCOCO testB | accuracy (IoU>0.5): 36.98 -> 38.59 | Hurt |
| add adaptive reconstruction to entity-enhanced model (ablation) | EARN · RefCOCOg val | accuracy (IoU>0.5): 38.37 -> 45.33 | Improved |
| use soft-context pooling instead of max-context pooling (ablation) | EARN · RefCOCOg val | accuracy (IoU>0.5): 42.81 -> 45.33 | Improved |
| use soft-context pooling instead of max-context pooling on complex relationships (ablation) | EARN · RefCOCOg val (complex) | accuracy (IoU>0.5): 43.11 -> 46.47 | Improved |
| use hard subject filter instead of soft subject filter (ablation) | EARN · RefCOCOg val | accuracy (IoU>0.5): 38.45 -> 40.24 | Improved |
| add location features to attribute features (ablation) | EARN · RefCOCO val | accuracy (IoU>0.5): 15.58 -> 36.47 | Improved |
| add context features (ablation) | EARN · RefCOCO+ val | accuracy (IoU>0.5): 20.33 -> 36.49 | Improved |
Closest papers by meaning
- Adaptive Reconstruction Network for Weakly Supervised Referring Expression Grounding (ICCV 2019 · similarity 0.96)
- Discriminative Triad Matching and Reconstruction for Weakly Referring Expression Grounding (TPAMI 2021 · similarity 0.91)
- Knowledge Aided Consistency for Weakly Supervised Phrase Grounding (CVPR 2018 · similarity 0.91)
- Ref-NMS: Breaking Proposal Bottlenecks in Two-Stage Referring Expression Grounding (AAAI 2021 · similarity 0.90)
- Using Syntax to Ground Referring Expressions in Natural Images (AAAI 2018 · similarity 0.90)
- Hierarchical Alignment-enhanced Adaptive Grounding Network for Generalized 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.
