Research map / Detection and segmentation / CNN features
CornerNet: Detecting Objects as Paired Keypoints
Hei Law, Jia Deng
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 corner pooling (ablation) | CornerNet · MS COCO val (5k) | AP: 36.5 -> 38.4 | Improved |
| add corner pooling (medium and large objects) (ablation) | CornerNet · MS COCO val (5k) | APm / APl: 38.9 -> 40.5; 49.4 -> 51.8 | Improved |
| add corner pooling to top-left corner detection, top-left quadrant (ablation) | CornerNet · MS COCO val corner mAP | mAP: 66.1 -> 69.2 | Improved |
| add corner pooling to bottom-right corner detection, bottom-right quadrant (ablation) | CornerNet · MS COCO val corner mAP | mAP: 65.0 -> 67.6 | Improved |
| reduce penalty near positives with fixed radius (ablation) | CornerNet · MS COCO val (5k) | AP: 32.9 -> 35.6 | Improved |
| use object-dependent radius instead of fixed radius (ablation) | CornerNet · MS COCO val (5k) | AP: 35.6 -> 38.4 | Improved |
| use hourglass backbone instead of FPN (ResNet-101) (ablation) | CornerNet · MS COCO val (5k) | AP: 30.2 -> 38.4 | Improved |
| detect corners instead of anchor boxes with hourglass backbone (ablation) | CornerNet · MS COCO val (5k) | AP: 32.9 -> 38.4 | Improved |
| replace predicted heatmaps with ground truth (ablation) | CornerNet · MS COCO val (5k) | AP: 38.4 -> 73.1 | Improved |
| also replace predicted offsets with ground truth (ablation) | CornerNet · MS COCO val (5k) | AP: 73.1 -> 86.1 | Improved |
| CornerNet vs Cascade R-CNN + IoU-Net at high IoU | CornerNet vs baselines · MS COCO (IoU 0.9) | AP90: 19.5 -> 23.4 | Improved |
| multi-scale testing with flip over single scale | CornerNet · MS COCO test-dev | AP: 40.6 -> 42.2 | Improved |
Later papers that compared with this method
13 later accepted papers report CornerNet as a baseline in their results tables; 13 of them report a better number on most of the results they share with it.
- CenterNet++ for Object Detection (TPAMI 2023 · better on 4 of 4 shared results)
- Cornerformer: Purifying Instances for Corner-Based Detectors (ECCV 2022 · better on 2 of 2 shared results)
- HoughNet: Integrating Near and Long-Range Evidence for Visual Detection (TPAMI 2022 · better on 1 of 1 shared result)
- Learning to Match Anchors for Visual Object Detection (TPAMI 2021 · better on 1 of 1 shared result)
- Corner Proposal Network for Anchor-free, Two-stage Object Detection (ECCV 2020 · better on 2 of 2 shared results)
- RepPoints v2: Verification Meets Regression for Object Detection (NeurIPS 2020 · better on 1 of 1 shared result)
- Dive Deeper Into Box for Object Detection (ECCV 2020 · better on 1 of 1 shared result)
- CenterNet: Keypoint Triplets for Object Detection (ICCV 2019 · better on 4 of 4 shared results)
5 more not listed.
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Cascade R-CNN — Cascade R-CNN: Delving Into High Quality Object Detection (CVPR 2018) · better on 1 of 2 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
- Corner Proposal Network for Anchor-free, Two-stage Object Detection (ECCV 2020 · similarity 0.90)
- CenterNet: Keypoint Triplets for Object Detection (ICCV 2019 · similarity 0.90)
- Gradient Corner Pooling for Keypoint-Based Object Detection (AAAI 2023 · similarity 0.90)
- CenterNet++ for Object Detection (TPAMI 2023 · similarity 0.88)
- Deep Corner (IJCV 2023 · similarity 0.88)
- Cornerformer: Purifying Instances for Corner-Based Detectors (ECCV 2022 · 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.
