Research map / Detection and segmentation / Open-set and few-shot detection
Multi-faceted Distillation of Base-Novel Commonality for Few-Shot Object Detection
Shuang Wu, Wenjie Pei, Dianwen Mei, Fanglin Chen, Jiandong Tian, Guangming Lu
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 recognition-related commonality distillation (ablation) | DeFRCN · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 58.5 -> 62.3 | Improved |
| add localization-related commonality distillation (ablation) | DeFRCN · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 58.5 -> 59.9 | Improved |
| add distribution commonality (ablation) | DeFRCN · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 58.5 -> 62.6 | Improved |
| combine all three commonalities (ablation) | DeFRCN · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 62.3 -> 63.4 | Improved |
| learn commonalities in the detector's own feature space instead of an independent ImageNet space (ablation) | DeFRCN · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 59.6 -> 62.3 | Improved |
| transfer only variance instead of mean and variance for distribution commonality (ablation) | DeFRCN · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 57.5 -> 62.6 | Improved |
| add MFDC to TFA w/cos | TFA w/cos · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 39.8 -> 45.2 | Improved |
| add MFDC to FSCE | FSCE · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 44.2 -> 48.0 | Improved |
| add MFDC to DeFRCN on COCO 1-shot | DeFRCN · MS COCO novel (1-shot) | nAP: 6.5 -> 10.8 | Improved |
| add commonality distillation | DeFRCN · VOC Novel Split 1 base classes (3-shot) | bAP50: 76.8 -> 78.3 | Improved |
| add commonality distillation | DeFRCN · VOC Novel Split 1 base classes (2-shot) | bAP50: 78.1 -> 78.0 | No clear effect |
| set loss weight lambda_c to 1.0 instead of 0.1 (ablation) | DeFRCN + recognition commonality · PASCAL VOC Novel Split 1 (1-shot) | nAP50: 62.3 -> 60.7 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- DeFRCN — DeFRCN: Decoupled Faster R-CNN for Few-Shot Object Detection (ICCV 2021) · better on 11 of 11 shared results
- FSCE — FSCE: Few-Shot Object Detection via Contrastive Proposal Encoding (CVPR 2021) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Few-Shot Object Detection by Knowledge Distillation Using Bag-of-Visual-Words Representations (ECCV 2022 · similarity 0.91)
- Few-Shot Object Detection via Association and DIscrimination (NeurIPS 2021 · similarity 0.90)
- Disentangle and Remerge: Interventional Knowledge Distillation for Few-Shot Object Detection from a Conditional Causal Perspective (AAAI 2023 · similarity 0.90)
- Structured Knowledge Distillation for Accurate and Efficient Object Detection (TPAMI 2023 · similarity 0.90)
- Learning Efficient Object Detection Models with Knowledge Distillation (NeurIPS 2017 · similarity 0.89)
- Improve Object Detection with Feature-based Knowledge Distillation: Towards Accurate and Efficient Detectors (ICLR 2021 · similarity 0.89)
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.
