Research map / Detection and segmentation / CNN features
Construct Effective Geometry Aware Feature Pyramid Network for Multi-Scale Object Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 FPN with GaFPN | Faster R-CNN with ResNet-50 · COCO test-dev | AP: 35.9 -> 37.8 | Helped |
| Replace FPN with GaFPN | Cascade R-CNN with ResNet-50 · COCO test-dev | AP: 40.2 -> 42.2 | Helped |
| Replace FPN with GaFPN | Dynamic R-CNN with ResNet-50 · COCO test-dev | AP: 38.3 -> 40.0 | Helped |
| Use GaFPN with SABL, without GPH | SABL with ResNet-50 · COCO test-dev | AP: 39.3 -> 40.6 | Helped |
| Add GaFPN to AugFPN | AugFPN with ResNet-50 · COCO test-dev | AP: 38.3 -> 39.1 | Helped |
| Replace FPN with GaFPN | Mask R-CNN with ResNet-50 · COCO test-dev | AP: 36.9 -> 38.5 | Helped |
| Add GMM (ablation) | Faster R-CNN FPN with ResNet-50 · COCO val2017 | AP: 35.6 -> 36.2 | Helped |
| Add GPH (ablation) | Faster R-CNN FPN with ResNet-50 · COCO val2017 | AP: 35.6 -> 36.1 | Helped |
| Add both GMM and GPH (ablation) | Faster R-CNN FPN with ResNet-50 · COCO val2017 | AP: 35.6 -> 36.9 | Helped |
| Generate mapping weights using proposal area and aspect ratio (ablation) | GMM · COCO val2017 | AP: 35.6 -> 36.2 | Helped |
| Add proposal coordinates to area and aspect-ratio inputs (ablation) | GMM · COCO val2017 | AP: 36.2 -> 36.2 | No clear effect |
| Use learned geometry-aware fusion instead of sum or max fusion (ablation) | GMM · COCO val2017 | AP: 35.8 -> 36.2 | Helped |
Closest papers by meaning
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.90)
- AugFPN: Improving Multi-Scale Feature Learning for Object Detection (CVPR 2020 · similarity 0.90)
- M2Det: A Single-Shot Object Detector Based on Multi-Level Feature Pyramid Network (AAAI 2019 · similarity 0.89)
- A2-FPN: Attention Aggregation Based Feature Pyramid Network for Instance Segmentation (CVPR 2021 · similarity 0.88)
- FAS-Net: Construct Effective Features Adaptively for Multi-Scale Object Detection (AAAI 2020 · similarity 0.88)
- HS-FPN: High Frequency and Spatial Perception FPN for Tiny Object Detection (AAAI 2025 · 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.