Research map / Detection and segmentation / Weakly supervised detection and segmentation
Interpolation-Based Semi-Supervised Learning for 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 ISD to CSD | SSD300 · PASCAL VOC07 test (VOC07 labeled, VOC12 unlabeled) | mAP (%): 72.3 -> 74.4 | Helped |
| add ISD to CSD | SSD300 · MS COCO test-dev | AP IoU 0.5:0.95: 19.8 -> 21.0 | Helped |
| add ISD to CSD | SSD300 · MS COCO test-dev | AP IoU 0.5: 35.8 -> 37.7 | Helped |
| add ISD to CSD | SSD300 · MS COCO test-dev | AP IoU 0.75: 19.8 -> 21.1 | Helped |
| add ISD losses in supervised training | SSD300 · PASCAL VOC07 test (VOC07 labeled only) | mAP (%): 70.2 -> 72.3 | Helped |
| add CSD regularization without ISD | SSD300 supervised training · PASCAL VOC07 test (VOC07 labeled only) | mAP (%): 70.2 -> 69.3 | Hurt |
| use Type-II rather than Type-I interpolation loss alone (ablation) | ISD with CSD · PASCAL VOC07 test | mAP (%): 71.9 -> 73.8 | Helped |
| add both Type-I and Type-II losses rather than Type-II alone (ablation) | ISD with CSD · PASCAL VOC07 test | mAP (%): 73.8 -> 74.1 | Helped |
| set mixing-distribution alpha to 100 rather than 1 (ablation) | ISD Type-I + Type-II with CSD · PASCAL VOC07 test | mAP (%): 72.9 -> 74.4 | Helped |
| set mixing-distribution alpha to 1000 rather than 100 (ablation) | ISD Type-I + Type-II with CSD · PASCAL VOC07 test | mAP (%): 74.4 -> 74.3 | Hurt |
| combine Type-II classification and localization losses rather than classification alone (ablation) | ISD Type-II with CSD · PASCAL VOC07 test | mAP (%): 74.0 -> 74.2 | Helped |
| add Type-II classification loss rather than using localization loss alone (ablation) | ISD Type-II with CSD · PASCAL VOC07 test | mAP (%): 73.1 -> 74.0 | Helped |
Closest papers by meaning
- Supervision Interpolation via LossMix: Generalizing Mixup for Object Detection and Beyond (AAAI 2024 · similarity 0.89)
- Semi-DETR: Semi-Supervised Object Detection With Detection Transformers (CVPR 2023 · similarity 0.88)
- Unbiased Teacher for Semi-Supervised Object Detection (ICLR 2021 · similarity 0.88)
- Scale-Equivalent Distillation for Semi-Supervised Object Detection (CVPR 2022 · similarity 0.87)
- Sparse Semi-DETR: Sparse Learnable Queries for Semi-Supervised Object Detection (CVPR 2024 · similarity 0.86)
- Relational Matching for Weakly Semi-Supervised Oriented Object Detection (CVPR 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.