Research map / Detection and segmentation / Open-set and few-shot detection
Rethinking Pseudo-Labels: Multi-Granularity Supervision for Domain Adaptive Object Detection
Haohao Ma, Weimin Lu, Qiqi Ge
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 image-level supervision (ablation) | AT (instance-level only) · Pascal VOC -> Clipart1k | mAP50: 45.7 -> 46.1 | Mixed |
| add instance-level supervision to image-level supervision (ablation) | MGPL image-level only · Pascal VOC -> Clipart1k | mAP50: 46.1 -> 50.6 | Improved |
| combine image-level and instance-level supervision (ablation) | AT · Pascal VOC -> Clipart1k | mAP50: 45.7 -> 50.6 | Improved |
| add coarse-to-fine collaborative filtering (ablation) | MGPL w/o Filter · Pascal VOC -> Clipart1k | mAP50: 50.6 -> 51.4 | Improved |
| increase lambda_img from 0.1 to 0.5 (ablation) | MGPL · Pascal VOC -> Clipart1k | mAP50: 48.5 -> 51.4 | Improved |
| increase lambda_img from 0.5 to 2.0 (ablation) | MGPL · Pascal VOC -> Clipart1k | mAP50: 51.4 -> 50.8 | Hurt |
| use MGPL instead of AT under fully flipped pseudo-box class labels | AT · Pascal VOC -> Clipart1k (noise ratio 1.0) | mAP50 gap: +5.7% -> +9.3% | Improved |
| inject full pseudo-label noise (flip all box class labels) | AT · Pascal VOC -> Clipart1k | mAP50: 45.7 -> 38.6 | Hurt |
| inject full pseudo-label noise (flip all box class labels) | MGPL w/o Filter · Pascal VOC -> Clipart1k | mAP50 change: -4.5% | Hurt |
| use MGPL instead of CMT | CMT · Cityscapes -> Foggy Cityscapes | mAP50: 51.9 -> 52.7 | Improved |
| use MGPL instead of CMT | CMT · KITTI -> Cityscapes (car) | AP50 car: 64.3 -> 64.3 | No clear effect |
Closest papers by meaning
- Robust Domain Adaptive Object Detection With Unified Multi-Granularity Alignment (TPAMI 2024 · similarity 0.90)
- Large Self-Supervised Models Bridge the Gap in Domain Adaptive Object Detection (CVPR 2025 · similarity 0.90)
- Enhancing Source-Free Domain Adaptive Object Detection with Low-confidence Pseudo Label Distillation (ECCV 2024 · similarity 0.89)
- Universal Domain Adaptive Object Detection via Dual Probabilistic Alignment (AAAI 2025 · similarity 0.88)
- CAT: Exploiting Inter-Class Dynamics for Domain Adaptive Object Detection (CVPR 2024 · similarity 0.88)
- Towards Unsupervised Model Selection for Domain Adaptive Object Detection (NeurIPS 2024 · 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.
