Research map / Detection and segmentation / Open-set and few-shot detection
Adaptive Hierarchical Representation Learning for Long-Tailed Object Detection
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 hierarchical coarse-to-fine representation learning with AHR loss | Baseline++ · LVIS v0.5 val | box AP: 26.7 -> 27.4 | Helped |
| Add hierarchical coarse-to-fine representation learning with AHR loss | Baseline++ · LVIS v0.5 val | segmentation AP: 26.8 -> 27.3 | Helped |
| Add hierarchical coarse-to-fine representation learning with AHR loss | Mask R-CNN · LVIS v1.0 val | box AP: 21.7 -> 28.7 | Helped |
| Add EQL loss (ablation) | Mask R-CNN · LVIS v0.5 val | segmentation AP: 24.2 -> 25.9 | Helped |
| Add GIoU regression loss (ablation) | Baseline++ with EQL · LVIS v0.5 val | box AP: 25.5 -> 25.8 | Helped |
| Add proposal oversampling for rare classes (ablation) | Baseline++ with EQL and GIoU · LVIS v0.5 val | box AP: 25.8 -> 26.1 | Helped |
| Add random cropping and color jitter (ablation) | Baseline++ with proposal oversampling · LVIS v0.5 val | segmentation AP: 26.4 -> 26.7 | Helped |
| Add cosine similarity classifier (ablation) | Baseline++ with data augmentation · LVIS v0.5 val | box AP: 26.5 -> 26.7 | Helped |
| Replace WordNet grouping with K-Means using 108 clusters (ablation) | AHRL · LVIS v0.5 val | segmentation AP: 26.9 -> 27.0 | Helped |
| Set K-Means cluster count to 200 rather than 108 (ablation) | AHRL · LVIS v0.5 val | box AP: 26.9 -> 27.4 | Helped |
| Set K-Means cluster count to 400 rather than 200 (ablation) | AHRL · LVIS v0.5 val | segmentation AP: 27.3 -> 26.8 | Hurt |
| Fine-tune the baseline++ model without AHR loss (ablation) | Baseline++ · LVIS v0.5 val | segmentation AP: 26.8 -> 27.0 | No clear effect |
Closest papers by meaning
- SimLTD: Simple Supervised and Semi-Supervised Long-Tailed Object Detection (CVPR 2025 · similarity 0.91)
- Adaptive Class Suppression Loss for Long-Tail Object Detection (CVPR 2021 · similarity 0.90)
- Long-tailed Object Detection Pretraining: Dynamic Rebalancing Contrastive Learning with Dual Reconstruction (NeurIPS 2024 · similarity 0.88)
- OFBD: Object-Focused Background Debiasing for Long-Tailed Learning (NeurIPS 2026 · similarity 0.87)
- Fractal Calibration for Long-tailed Object Detection (CVPR 2025 · similarity 0.87)
- Pursuing Better Decision Boundaries for Long-Tailed Object Detection via Category Information Amount (ICLR 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.