Research map / Detection and segmentation / CNN features
Scale Normalized Image Pyramids with AutoFocus for Object Detection
Bharat Singh, Mahyar Najibi, Abhishek Sharma, Larry Steven Davis
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 |
|---|---|---|---|
| use SNIP scale normalization instead of multi-scale train/test | R-FCN ResNet-50 · COCO | AP: 35.6 -> 37.8 | Improved |
| train on multi-scale without scale normalization (MST) instead of multi-scale test only (ablation) | R-FCN ResNet-50 · COCO | AP: 35.9 -> 35.6 | Hurt |
| restrict training to objects <80px at 1400 resolution (ablation) | R-FCN · COCO small objects | mAP: 19.6 -> 16.4 | Hurt |
| train at 1400x2000 instead of 800x1400 resolution | R-FCN · COCO small objects | mAP: 19.6 -> 19.9 | No clear effect |
| add SNIP to RPN training (ablation) | RPN ResNet-50 · COCO | AR: 61.3 -> 64.0 | Improved |
| add SNIP to RPN and RCN (ablation) | D-R-FCN DPN-98 · COCO | AP: 41.2 -> 44.7 | Improved |
| add SNIP to RPN and RCN (ablation) | Faster-RCNN ResNet-101 · COCO | AP: 42.6 -> 44.4 | Improved |
| add negative chip mining (ablation) | SNIPER · COCO | AP: 43.4 -> 46.1 | Improved |
| add negative chip sampling (ablation) | SNIPER RPN · COCO | AR: 65.4 -> 65.4 | No clear effect |
| add third (highest resolution) scale (ablation) | SNIPER · COCO | AP: 43.3 -> 46.1 | Improved |
| use SNIPER chips instead of full image pyramid (no batch norm) | SNIP · COCO | AP: 43.6 -> 43.5 | No clear effect |
| use SNIPER with image pyramids instead of SSD | MobileNet-v2 · COCO | AP: 22.1 -> 34.5 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SNIPER — SNIPER: Efficient Multi-Scale Training (NeurIPS 2018) · better on 1 of 1 shared result
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
- An Analysis of Scale Invariance in Object Detection SNIP (CVPR 2018 · similarity 0.92)
- Scale-Equalizing Pyramid Convolution for Object Detection (CVPR 2020 · similarity 0.90)
- POD: Practical Object Detection With Scale-Sensitive Network (ICCV 2019 · similarity 0.89)
- M2Det: A Single-Shot Object Detector Based on Multi-Level Feature Pyramid Network (AAAI 2019 · similarity 0.88)
- Learning Rich Features at High-Speed for Single-Shot Object Detection (ICCV 2019 · similarity 0.88)
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.88)
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.
