Research map / Detection and segmentation / Open-set and few-shot detection
Self-Training and Adversarial Background Regularization for Unsupervised Domain Adaptive One-Stage Object Detection
Seunghyeon Kim, Jaehoon Choi, Taekyung Kim, Changick Kim
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 weak negative mining (Weak Mask) instead of Mask (ablation) | self-training w/o SRRS · VOC -> Clipart1k | mAP: 16.8 -> 31.3 | Improved |
| add weak negative mining (Weak Mask) instead of Mask (ablation) | self-training with SRRS · VOC -> Clipart1k | mAP: 29.2 -> 33.8 | Improved |
| add SRRS to Mask (ablation) | Mask self-training · VOC -> Clipart1k | mAP: 16.8 -> 29.2 | Improved |
| add SRRS to Weak Mask (ablation) | Weak Mask self-training · VOC -> Clipart1k | mAP: 31.3 -> 33.8 | Improved |
| add SRRS alone to naive self-training (ablation) | ST · VOC -> Clipart1k | mAP: 10.7 -> 10.5 | No clear effect |
| add Mask (no negatives) to naive self-training (ablation) | ST · VOC -> Clipart1k | mAP: 10.7 -> 16.8 | Improved |
| use naive self-training with pseudo-labels as ground truth (ablation) | Base SSD300 · VOC -> Clipart1k | mAP: 26.7 -> 10.7 | Hurt |
| add BSR to base network | Base SSD300 · VOC -> Clipart1k | mAP: 26.7 -> 34.0 | Improved |
| add WST on top of BSR | BSR · VOC -> Clipart1k | mAP: 34.0 -> 35.7 | Improved |
| add BSR and WST | Base SSD300 · VOC -> Comic2k | mAP: 21.9 -> 26.8 | Improved |
| add BSR and WST | Base SSD300 · VOC -> Watercolor2k | mAP: 47.1 -> 49.9 | Improved |
| use DANN instead of base | Base SSD300 · VOC -> Watercolor2k | mAP: 47.1 -> 41.5 | Hurt |
Closest papers by meaning
- Cross-Domain Adaptive Teacher for Object Detection (CVPR 2022 · similarity 0.91)
- A Free Lunch for Unsupervised Domain Adaptive Object Detection without Source Data (AAAI 2021 · similarity 0.91)
- Diversify and Match: A Domain Adaptive Representation Learning Paradigm for Object Detection (CVPR 2019 · similarity 0.90)
- Enhancing Source-Free Domain Adaptive Object Detection with Low-confidence Pseudo Label Distillation (ECCV 2024 · similarity 0.90)
- Simplifying Source-Free Domain Adaptation for Object Detection: Effective Self-Training Strategies and Performance Insights (ECCV 2024 · similarity 0.90)
- SSAL: Synergizing between Self-Training and Adversarial Learning for Domain Adaptive Object Detection (NeurIPS 2021 · similarity 0.90)
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.
