Research map / Detection and segmentation / CNN features
G-CNN: An Iterative Grid Based Object Detector
Check 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 selective-search proposals with an approximately 180-box grid and iterative box regression | Fast R-CNN detector with VGG16 · VOC 2007 test | mAP: 66.9 -> 66.8 | No clear effect |
| Replace selective-search proposals with an approximately 180-box grid and iterative box regression | Fast R-CNN detector with AlexNet · VOC 2007 test | mAP: 57.1 -> 57.2 | Helped |
| Use 5 test iterations and reduced grid overlaps instead of 3 iterations with training overlaps (ablation) | G-CNN with AlexNet · VOC 2007 test | mAP: 56.7 -> 57.2 | Helped |
| Replace stepwise training with one-step training on tuples from all steps (ablation) | G-CNN with AlexNet · VOC 2007 test | mAP: 57.2 -> 54.5 | Hurt |
| Apply the Fast R-CNN regressor iteratively to grid boxes instead of using G-CNN's stepwise-trained regressor (ablation) | Grid-based detection with AlexNet · VOC 2007 test | mAP: 57.2 -> 54.1 | Hurt |
| Remove the proposal-generation stage and use iterative grid regression | Fast R-CNN with VGG16 · Detection runtime per image | runtime: 2050 ms/im -> 363 ms/im; 5X faster | Helped |
| Train on VOC 2007 and VOC 2012 trainval rather than VOC 2012 trainval alone (ablation) | G-CNN with VGG16 · VOC 2012 test | mAP: 65.5 -> 66.4* | Helped |
Closest papers by meaning
- Sparse R-CNN: End-to-End Object Detection With Learnable Proposals (CVPR 2021 · similarity 0.88)
- Grid R-CNN (CVPR 2019 · similarity 0.88)
- Group R-CNN for Weakly Semi-Supervised Object Detection With Points (CVPR 2022 · similarity 0.87)
- R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016 · similarity 0.87)
- G-NAS: Generalizable Neural Architecture Search for Single Domain Generalization Object Detection (AAAI 2024 · similarity 0.86)
- TGBFormer: Transformer-GraphFormer Blender Network for Video 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.