Research map / Detection and segmentation / CNN features
GiraffeDet: A Heavy-Neck Paradigm for 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 the conventional backbone/neck allocation with a lightweight S2D chain and heavy GFPN | GiraffeDet · COCO test-dev | AP: 44.8 -> 45.6 | Helped |
| Add multi-scale testing (ablation) | GiraffeDet-D29 · COCO test-dev | AP: 51.3 -> 54.1 | Helped |
| Add log₂n skip-layer connections (ablation) | S2D chain + GFPN · COCO val-2017, s-3x | AP: 41.2 -> 41.8 | Helped |
| Replace dense-link with log₂n-link (ablation) | S2D chain + GFPN · COCO val-2017, s-3x | AP: 41.3 -> 41.8 | Helped |
| Replace stacked FPN with GFPN-log₂n (ablation) | S2D chain detector · COCO val-2017, s-3x | AP: 38.8 -> 41.8 | Helped |
| Replace stacked PANet with GFPN-log₂n (ablation) | S2D chain detector · COCO val-2017, s-3x | AP: 40.5 -> 41.8 | Helped |
| Replace stacked BiFPN with GFPN-log₂n (ablation) | S2D chain detector · COCO val-2017, s-3x | AP: 41.0 -> 41.8 | Helped |
| Add synchronized batch normalization (ablation) | S2D chain + GFPN-D11 · COCO val-2017, s-3x | AP: 41.8 -> 42.9 | Helped |
| Add DCN and synchronized batch normalization (ablation) | S2D chain + GFPN-D11 · COCO val-2017, s-3x | AP: 42.9 -> 45.3 | Helped |
| Use GFPN-D11 instead of GFPN-tiny (ablation) | Res2Net-101-DCN + GFLV2 · COCO val-2017, p-2x | AP: 50.2 -> 51.1 | Helped |
| Replace FPN with GFPN-D11 (ablation) | Res2Net-101-DCN + GFLV2 · COCO val-2017, p-2x | AP: 49.9 -> 51.1 | Helped |
| Use concatenation instead of summation for feature fusion (ablation) | GiraffeDet feature fusion · COCO | AP | Helped |
Closest papers by meaning
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.87)
- Unbalanced Optimal Transport: A Unified Framework for Object Detection (CVPR 2023 · similarity 0.87)
- R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016 · similarity 0.87)
- M2Det: A Single-Shot Object Detector Based on Multi-Level Feature Pyramid Network (AAAI 2019 · similarity 0.87)
- SimLTD: Simple Supervised and Semi-Supervised Long-Tailed Object Detection (CVPR 2025 · similarity 0.85)
- Unsupervised Object Detection with Theoretical Guarantees (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.