Research map / Label-efficient and robust learning / Recognition and long-tail
Category-Specific Selective Feature Enhancement for Long-Tailed Multi-Label Image Classification
Ruiqi Du, Xu Tang, Xiangrong Zhang, Jingjing Ma
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 progressive attention enhancement to Bilateral-TFS (ablation) | Bilateral-TFS · LT-VOC | total mAP: 81.58 -> 85.66 | Improved |
| add progressive attention enhancement to Bilateral-TFS (ablation) | Bilateral-TFS · LT-COCO | total mAP: 56.38 -> 65.79 | Improved |
| add progressive attention enhancement to CAE-Net (ablation) | CAE-Net · LT-VOC | total mAP: 81.61 -> 85.83 | Improved |
| add progressive attention enhancement to CAE-Net (ablation) | CAE-Net · LT-COCO | total mAP: 57.64 -> 66.18 | Improved |
| add progressive attention enhancement to CPRFL (ablation) | CPRFL · LT-VOC | total mAP: 86.28 -> 86.32 | Improved |
| add progressive attention enhancement to CPRFL (ablation) | CPRFL · LT-COCO | tail mAP: 61.33 -> 61.48 | Improved |
| CSSFE versus second-best CPRFL | CSSFE · LT-VOC | total/head/medium/tail mAP: +0.10%/0.09%/0.65%/0.83% | Improved |
| CSSFE versus second-best CPRFL | CSSFE · LT-COCO | total/head/medium/tail mAP: +0.24%/0.61%/0.22%/1.08% | Improved |
| add CLIP text label embeddings to baseline (Net-1 vs Net-0) (ablation) | CSSFE baseline · LT-VOC and LT-COCO | total mAP | Improved |
| add unmasked semantic attention adjustment (Net-2 vs Net-1) (ablation) | CSSFE · LT-VOC and LT-COCO | head/tail mAP | Mixed |
| add one-way head-to-medium-to-tail binary mask (Net-3 vs Net-2) (ablation) | CSSFE · LT-VOC and LT-COCO | total/head/medium/tail mAP | Improved |
| DB Focal re-weighting instead of BBN re-sampling | long-tailed baselines · LT-COCO | head vs tail mAP: head gap 1.34%, tail gap 6.12% | Improved |
Closest papers by meaning
- Long-Tailed Multi-Label Visual Recognition by Collaborative Training on Uniform and Re-Balanced Samplings (CVPR 2021 · similarity 0.88)
- Self Supervision to Distillation for Long-Tailed Visual Recognition (ICCV 2021 · similarity 0.88)
- MLC-NC: Long-Tailed Multi-Label Image Classification Through the Lens of Neural Collapse (AAAI 2025 · similarity 0.87)
- Continuous Contrastive Learning for Long-Tailed Semi-Supervised Recognition (NeurIPS 2024 · similarity 0.86)
- Correlative and Discriminative Label Grouping for Multi-Label Visual Prompt Tuning (CVPR 2025 · similarity 0.85)
- Double-Layer Hybrid-Label Identification Feature Selection for Multi-View Multi-Label Learning (AAAI 2024 · similarity 0.85)
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.
