Research map / Detection and segmentation / CNN features
HoughNet: Integrating Near and Long-Range Evidence for Visual Detection
Nermin Samet, Samet Hiçsönmez, Emre Akbaş
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 voting module | OAP R-101-DCN · COCO val2017 (SS) | AP: 36.2 -> 37.2 | Improved |
| add voting module | OAP HG-104 · COCO val2017 (MS) | AP: 45.1 -> 46.1 | Improved |
| replace dilated convolution with voting module (ablation) | OAP R-101-DCN · COCO val2017 | AP: 36.6 -> 37.3 | Improved |
| add outer ring votes (vs only center) (ablation) | HoughNet · COCO minitrain | AP: 23.8 -> 24.6 | Improved |
| add center ring (vs no center) (ablation) | HoughNet · COCO minitrain | AP: 24.4 -> 24.6 | Improved |
| use 3 rings instead of 5 (ablation) | HoughNet · COCO minitrain | AP: 24.6 -> 24.8 | Improved |
| use 90 degree angle bins vs 180 (ablation) | HoughNet · COCO minitrain | AP50: 41.1 -> 41.5 | Improved |
| initialize from ExtremeNet weights | HoughNet HG-104 · COCO test-dev (SS) | AP: 40.8 -> 43.1 | Improved |
| scalable voting with N=8 shared tensors vs N=4 (ablation) | HoughNet · COCO minitrain | AP: 24.9 -> 25.9 | Improved |
| scalable voting instead of per-class evidence (ablation) | HoughNet R-101-DCN · COCO val2017 | AP: 14.3 FPS -> 26.3 FPS; 37.3 -> 36.3 AP | Mixed |
| add temporal voting with motion features | HoughNet · ImageNet VID | mAP: 68.8 -> 74.9 | Improved |
| add voting module to instance segmentation | OAP · COCO val2017 | AP seg: 27.2 -> 28.4 | Improved |
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.
- CornerNet — CornerNet: Detecting Objects as Paired Keypoints (ECCV 2018) · better on 1 of 1 shared result
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
- HoughNet: Integrating near and long-range evidence for bottom-up object detection (ECCV 2020 · similarity 0.98)
- ObjectNet: A large-scale bias-controlled dataset for pushing the limits of object recognition models (NeurIPS 2019 · similarity 0.90)
- DeNet: Scalable Real-Time Object Detection With Directed Sparse Sampling (ICCV 2017 · similarity 0.89)
- Learning Rich Features at High-Speed for Single-Shot Object Detection (ICCV 2019 · similarity 0.88)
- Towards Large-Scale Small Object Detection: Survey and Benchmarks (TPAMI 2023 · similarity 0.88)
- SimLTD: Simple Supervised and Semi-Supervised Long-Tailed Object Detection (CVPR 2025 · similarity 0.87)
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.
