Research map / Detection and segmentation / CNN features
Parallel Feature Pyramid Network for Object Detection
Seung-Wook Kim, Hyong-Keun Kook, Jee-Young Sun, Mun-Cheon Kang, Sung-Jea Ko
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 SPP-based wide feature pyramid (ablation) | SSD baseline (320x320) · PASCAL VOC 2007 test | mAP (%): 76.2 -> 77.9 | Improved |
| add MSCA module to wide FP (ablation) | PFPNet without ARM · PASCAL VOC 2007 test | mAP (%): 77.9 -> 78.5 | Improved |
| add ARM anchor refinement (ablation) | PFPNet-R320 · PASCAL VOC 2007 test | mAP (%): 78.5 -> 80.7 | Improved |
| add ARM anchor refinement (ablation) | RefineDet320 · PASCAL VOC 2007 test | mAP (%): 77.3 -> 80.0 | Improved |
| add ARM anchor refinement (ablation) | FPN320 · PASCAL VOC 2007 test | mAP (%): 77.6 -> 79.6 | Improved |
| use 4 pyramid levels instead of 2 (ablation) | PFPNet (C_L=256) · PASCAL VOC 2007 test | mAP (%): 79.08 -> 80.72 | Improved |
| use 4 pyramid levels instead of 6 (ablation) | PFPNet (C_L=256) · PASCAL VOC 2007 test | mAP (%): 80.15 -> 80.72 | Improved |
| increase reduced channels C_L from 64 to 256 (ablation) | PFPNet (N=4) · PASCAL VOC 2007 test | mAP (%): 80.16 -> 80.72 | Improved |
| add skip connection from F_H in MSCA (ablation) | PFPNet · PASCAL VOC 2007 test | mAP (%) | Improved |
| replace SSD with PFPNet-R320 (same VGG-16, ARM) | RefineDet320 · MS COCO test-dev | AP50:95: 29.4 -> 31.8 | Improved |
| use multi-scale testing | PFPNet-R512 · MS COCO test-dev | AP50:95: 35.2 -> 39.4 | Improved |
| use FPN-style hourglass FP instead of ARM-free RefineDet FP | RefineDet320 vs FPN320 (no ARM) · PASCAL VOC 2007 test | mAP (%): 77.3 -> 77.6 | Improved |
Closest papers by meaning
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.95)
- M2Det: A Single-Shot Object Detector Based on Multi-Level Feature Pyramid Network (AAAI 2019 · similarity 0.91)
- GraphFPN: Graph Feature Pyramid Network for Object Detection (ICCV 2021 · similarity 0.90)
- NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection (CVPR 2019 · similarity 0.90)
- Object Detection Networks on Convolutional Feature Maps (TPAMI 2016 · similarity 0.89)
- Construct Effective Geometry Aware Feature Pyramid Network for Multi-Scale Object Detection (AAAI 2022 · similarity 0.89)
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.
