Research map / Detection and segmentation / CNN features
Better to Follow, Follow to Be Better: Towards Precise Supervision of Feature Super-Resolution for Small Object Detection
Junhyug Noh, Wonho Bae, Wonhee Lee, Jinhwan Seo, Gunhee 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 content loss with SR target extractor supervision (ablation) | Faster R-CNN ResNet-50 + SR · Tsinghua-Tencent 100K | overall F1: 87.5 -> 89.7 | Improved |
| use SR target extractor (matched receptive field) instead of naive targets (ablation) | Faster R-CNN ResNet-50 + SR · Tsinghua-Tencent 100K | overall F1: 85.3 -> 89.7 | Improved |
| add naive supervision from base extractor (ablation) | Faster R-CNN ResNet-50 + SR · Tsinghua-Tencent 100K | overall F1: 87.5 -> 85.3 | Hurt |
| add feature-level SR for small proposals | Faster R-CNN ResNet-50 · Tsinghua-Tencent 100K small objects | F1: 74.9 -> 82.2 | Improved |
| add feature-level SR for small proposals | Faster R-CNN ResNet-101 · Tsinghua-Tencent 100K small objects | F1: 75.2 -> 84.3 | Improved |
| add feature-level SR for small proposals | Faster R-CNN MobileNet · Tsinghua-Tencent 100K small objects | F1: 63.4 -> 71.0 | Improved |
| add feature-level SR for small proposals | Faster R-CNN ResNet-101 · Tsinghua-Tencent 100K large objects | F1: 92.2 -> 93.2 | Improved |
| add feature-level SR for small proposals | Faster R-CNN MobileNet · Tsinghua-Tencent 100K large objects | F1: 87.1 -> 86.5 | Hurt |
| add feature-level SR for small proposals | Faster R-CNN ResNet-101 · COCO 2017 test-dev | AP-S: 11.3 -> 16.2 | Improved |
| add feature-level SR for small proposals | Faster R-CNN ResNet-50 · PASCAL VOC 2007 test | AP-S: 6.8 -> 10.5 | Improved |
| add feature-level SR for small proposals | Faster R-CNN MobileNet · COCO 2017 test-dev | AP-L: 29.2 -> 29.0 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Faster R-CNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · better on 17 of 17 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
- Feature Super-Resolution: Make Machine See More Clearly (CVPR 2018 · similarity 0.89)
- Deep Learning for Image Super-Resolution: A Survey (TPAMI 2020 · similarity 0.87)
- SOD-MTGAN: Small Object Detection via Multi-Task Generative Adversarial Network (ECCV 2018 · similarity 0.87)
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks (CVPR 2016 · similarity 0.87)
- Sparse R-CNN: An End-to-End Framework for Object Detection (TPAMI 2023 · similarity 0.87)
- Beyond Image Super-Resolution for Image Recognition with Task-Driven Perceptual Loss (CVPR 2024 · similarity 0.86)
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.
