Research map / Detection and segmentation / CNN features
Hyper-YOLO: When Visual Object Detection Meets Hypergraph Computation
Yifan Feng, Jiangang Huang, Shaoyi Du, Shihui Ying, Jun‐Hai Yong, Yipeng Li, Guiguang Ding, Rongrong Ji, Yue Gao
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 |
|---|---|---|---|
| replace C2f with MANet block (ablation) | Hyper-YOLO-S backbone · COCO val2017 | AP val: 44.9 -> 46.4 | Improved |
| use kernels [3,5,5,3] instead of all 3 (ablation) | MANet backbone · COCO val2017 | AP val: 46.3 -> 46.4 | Improved |
| use all-5 kernels instead of [3,5,5,3] (ablation) | MANet backbone · COCO val2017 | AP val: 46.4 -> 46.6 | No clear effect |
| add low-order graph learning in the neck (ablation) | Hyper-YOLO-S · COCO val2017 | AP val: 46.4 -> 47.6 | Improved |
| use high-order hypergraph instead of low-order graph learning (ablation) | Hyper-YOLO-S · COCO val2017 | AP val: 47.6 -> 48.0 | Improved |
| increase semantic collecting inputs from 3 to 5 levels (ablation) | Hyper-YOLO-S · COCO val2017 | AP val: 47.5 -> 48.0 | Improved |
| distance threshold 8 instead of 6 (ablation) | hypergraph construction · COCO val2017 | AP val: 47.6 -> 48.0 | Improved |
| distance threshold 8 instead of 10 (ablation) | hypergraph construction · COCO val2017 | AP val: 47.7 -> 48.0 | Improved |
| add hypergraph computation vs none (ablation) | Hyper-YOLO-S · COCO val2017 | AP val: 46.3 -> 48.0 | Improved |
| add backbone enhancement to YOLOv8-N (ablation) | Hyper-YOLO-N · COCO val2017 | AP val: 37.3% -> 39.9% | Improved |
| add neck enhancement after backbone (ablation) | Hyper-YOLO-N · COCO val2017 | AP val: 39.9% -> 41.8% | Improved |
| add backbone enhancement to YOLOv8-L (ablation) | Hyper-YOLO-L · COCO val2017 | AP val: 52.9% -> 53.0% | No clear effect |
Closest papers by meaning
- HyperNet: Towards Accurate Region Proposal Generation and Joint Object Detection (CVPR 2016 · similarity 0.89)
- YOLO9000: Better, Faster, Stronger (CVPR 2017 · similarity 0.88)
- SoftHGNN: Soft Hypergraph Neural Networks for General Visual Recognition (IJCV 2026 · similarity 0.88)
- Gold-YOLO: Efficient Object Detector via Gather-and-Distribute Mechanism (NeurIPS 2023 · similarity 0.87)
- YOLO-RD: Introducing Relevant and Compact Explicit Knowledge to YOLO by Retriever-Dictionary (ICLR 2025 · similarity 0.86)
- YOLOv10: Real-Time End-to-End Object Detection (NeurIPS 2024 · similarity 0.85)
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.
