Research map / Detection and segmentation / Open-set and few-shot detection
Meta R-CNN: Towards General Solver for Instance-Level Low-Shot Learning
Xiaopeng Yan, Ziliang Chen, Anni Xu, Xiaoxi Wang, Xiaodan Liang, Liang Lin
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 meta-learning (predictor-head attention vectors) (ablation) | Meta R-CNN · VOC2007 novel split 1, 3-shot | mAP: 9.0 -> 35.0 | Improved |
| add meta-learning (ablation) | Meta R-CNN · VOC2007 novel split 1, 10-shot | mAP: 40.5 -> 51.5 | Improved |
| meta-learn on RoI features instead of full image (ablation) | Meta R-CNN · VOC2007 novel split 1, 3-shot | mAP: 8.1 -> 35.0 | Improved |
| meta-learn on RoI features instead of full image (ablation) | Meta R-CNN · VOC2007 novel split 1, 10-shot | mAP: 32.0 -> 51.5 | Improved |
| add meta-loss diversifying attentive vectors (ablation) | Meta R-CNN · VOC2007 novel split 1, 3-shot | mAP: 24.2 -> 35.0 | Improved |
| add meta-loss diversifying attentive vectors (ablation) | Meta R-CNN · VOC2007 novel split 1, 10-shot | mAP: 46.6 -> 51.5 | Improved |
| use Meta R-CNN instead of FRCN+ft-full with ResNet-34 | Faster R-CNN ResNet-34 · VOC2007 novel split 1, 3-shot | mAP: 19.6 -> 25.3 | Improved |
| use Meta R-CNN instead of FRCN+ft-full with ResNet-101 | Faster R-CNN ResNet-101 · VOC2007 novel split 1, 10-shot | mAP: 45.6 -> 51.5 | Improved |
| use Meta R-CNN instead of FRCN+ft-full | Meta R-CNN · COCO minival 10-shot novel classes | AP: 6.5 -> 8.7 | Improved |
| use Meta R-CNN instead of FRCN+ft-full | Meta R-CNN · COCO minival 30-shot novel classes | AP: 11.1 -> 12.4 | Improved |
| use Meta R-CNN instead of YOLO-Few-shot | Meta R-CNN · VOC2007 novel split 3, 1-shot | mAP: 19.2 -> 14.3 | Hurt |
| apply PRN to Mask R-CNN | Mask R-CNN · COCO minival 20-shot box AP | AP: 4.5 -> 6.2 | Improved |
Closest papers by meaning
- Meta Faster R-CNN: Towards Accurate Few-Shot Object Detection with Attentive Feature Alignment (AAAI 2022 · similarity 0.91)
- Sparse R-CNN: End-to-End Object Detection With Learnable Proposals (CVPR 2021 · similarity 0.91)
- Meta-Learning to Detect Rare Objects (ICCV 2019 · similarity 0.90)
- Few-Shot Object Detection via Feature Reweighting (ICCV 2019 · similarity 0.90)
- Sparse R-CNN: An End-to-End Framework for Object Detection (TPAMI 2023 · similarity 0.90)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.88)
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.
