Research map / Detection and segmentation / CNN features
Training Multi-Object Detector by Estimating Bounding Box Distribution for Input Image
Jaeyoung Yoo, Hojun Lee, Inseop Chung, Geonseok Seo, Nojun Kwak
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 |
|---|---|---|---|
| use Cauchy instead of Gaussian distribution (ablation) | MDOD · COCO val2017 | AP: 32.7 -> 33.8 | Improved |
| add level-scale operation (ablation) | MDOD · COCO val2017 | AP: 32.9 -> 33.8 | Improved |
| add center-limit operation (on top of ltrb) (ablation) | MDOD · COCO val2017 | AP: 32.3 -> 32.9 | Improved |
| use ltrb-transformation instead of xywh (ablation) | MDOD · COCO val2017 | AP: 32.8 -> 33.8 | Improved |
| increase N_roi from N_gt x1 to x5 (ablation) | MDOD · COCO val2017 | AP: 33.8 -> 33.9 | No clear effect |
| replace conventional anchor-based training with density-estimation training | ResNet50-FPN 320x320 · COCO test-dev | AP: 30.1 -> 33.9 | Improved |
| replace conventional anchor-based training with density-estimation training | ResNet101-FPN 512x512 · COCO test-dev | AP: 36.6 -> 40.0 | Improved |
| use MDOD training instead of EfficientDet training | Efficient-D1 640x640 · COCO test-dev | AP: 39.6 -> 40.5 | Improved |
| use MDOD training instead of baseline | ResNet50-FPN 512x512 · COCO test-dev | FPS: 37.5 -> 43.5 | Improved |
| use MDOD on variable-size input (short-800) versus ATSS | ResNet-101 FPN · COCO test-dev short-800 | AP: 43.6 -> 42.2 | Hurt |
Closest papers by meaning
- End-to-End Multi-Object Detection with a Regularized Mixture Model (ICML 2023 · similarity 0.86)
- Active Learning for Deep Object Detection via Probabilistic Modeling (ICCV 2021 · similarity 0.86)
- Mixture Dense Regression for Object Detection and Human Pose Estimation (CVPR 2020 · similarity 0.86)
- Point-to-Box Network for Accurate Object Detection via Single Point Supervision (ECCV 2022 · similarity 0.84)
- Bounding Box Regression With Uncertainty for Accurate Object Detection (CVPR 2019 · similarity 0.84)
- Open-Vocabulary Object Detection Using Captions (CVPR 2021 · similarity 0.84)
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.
