Research map / Detection and segmentation / CNN features
SAN: Learning Relationship between Convolutional Features for Multi-Scale Object Detection
Yonghyun Kim, Bong-Nam Kang, Daijin Kim
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 SAN to R-FCN | R-FCN ResNet-101 · VOC07 test | mAP: 79.4 -> 80.6 | Improved |
| add SAN to Deformable R-FCN | Deformable R-FCN · MS COCO minival | mAP: 41.2 -> 43.3 | Improved |
| add SAN to Deformable R-FCN | Deformable R-FCN · VOC07 test | mAP: 82.1 -> 82.8 | Improved |
| add SAN to Faster RCNN | Faster RCNN · VOC07 test | mAP: 79.3 -> 79.9 | Improved |
| add scale-aware loss to SAN (ablation) | R-FCN SAN · VOC07 test | mAP: 80.1 -> 80.6 | Improved |
| add scale-aware loss to SAN (ablation) | Deformable R-FCN SAN · VOC07 test | mAP: 82.4 -> 82.8 | Improved |
| add SAN to R-FCN, small objects | R-FCN · MS COCO minival | mAP small: 16.4 -> 16.7 | Improved |
| add SAN to Deformable R-FCN, large objects | Deformable R-FCN · MS COCO minival | mAP large: 61.2 -> 64.6 | Improved |
| use identity instead of gaussian initialization of SAN (ablation) | SAN 3 partitions · VOC07 test | mAP: 79.97 -> 80.57 | Improved |
| use average instead of max pooling for SAN learning (ablation) | SAN 3 partitions · VOC07 test | mAP: 80.35 -> 80.57 | Improved |
| use 3 partitions instead of 2 (reference 224^2) (ablation) | SAN · VOC07 test | mAP: 80.44 -> 80.57 | Improved |
| use 3 partitions with split at 112^2/224^2 instead of 160^2/288^2 (ablation) | SAN · VOC07 test | mAP: 80.57 -> 80.16 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- R-FCN — R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016) · better on 6 of 6 shared results
- Faster RCNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · better on 2 of 2 shared results
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
- A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection (ECCV 2016 · similarity 0.89)
- An Analysis of Scale Invariance in Object Detection SNIP (CVPR 2018 · similarity 0.89)
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.88)
- POD: Practical Object Detection With Scale-Sensitive Network (ICCV 2019 · similarity 0.88)
- SCNN: A General Distribution Based Statistical Convolutional Neural Network with Application to Video Object Detection (AAAI 2019 · similarity 0.87)
- ScaleKD: Distilling Scale-Aware Knowledge in Small Object Detector (CVPR 2023 · 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.
