Research map / Detection and segmentation / CNN features
Mimicking Very Efficient Network for Object Detection
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 |
|---|---|---|---|
| Use proposal-sampled feature mimic instead of training from scratch | 1/2-Inception detector · Caltech test | MR−2: 24.63 -> 7.55 | Helped |
| Use feature mimic instead of ImageNet fine-tuning | 1/2-Inception detector · Caltech test | MR−2: 8.88 -> 7.55 | Helped |
| Use proposal-sampled feature mimic instead of training from scratch | 1/4-Inception detector · Caltech test | MR−2: 30.36 -> 10.02 | Helped |
| Add second-stage logits mimic (ablation) | Caltech Inception mimic detector · Caltech test | MR−2: 7.55 -> 7.31 | Helped |
| Add second-stage logits mimic (ablation) | Caltech 1/4-Inception mimic detector · Caltech test | MR−2: 10.02 -> 9.75 | Helped |
| Add second-stage logits mimic (ablation) | Caltech 1/8-Inception mimic detector · Caltech test | MR−2: 16.86 -> 15.32 | Helped |
| Add deconvolutional upsampling without feature mimic (ablation) | Low-resolution Inception R-FCN · Caltech test, 480×640 | MR−2: 14.64 -> 15.28 | Hurt |
| Add cross-scale feature mimic to the upsampled low-resolution network (ablation) | Low-resolution Inception R-FCN · Caltech test, 480×640 | MR−2: 15.28 -> 11.14 | Helped |
| Train with feature mimic rather than from scratch | 1/2-Inception Faster R-CNN · PASCAL VOC 2007 test | mAP: 49.21 -> 72.79 | Helped |
| Train with feature mimic rather than ImageNet fine-tuning | 1/2-Inception Faster R-CNN · PASCAL VOC 2007 test | mAP: 72.37 -> 72.79 | Helped |
| Add second-stage logits mimic (ablation) | PASCAL VOC 1/4-Inception detector · PASCAL VOC 2007 test | mAP: 65.76 -> 67.66 | Helped |
| Mimic at Inception-4d rather than Inception-3b (ablation) | 1/4-Inception Faster R-CNN · PASCAL VOC 2007 test | mAP: 42.46 -> 65.76 | Helped |
Closest papers by meaning
- Training-Time-Friendly Network for Real-Time Object Detection (AAAI 2020 · similarity 0.88)
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.88)
- Rethinking the Inception Architecture for Computer Vision (CVPR 2016 · similarity 0.87)
- R-FCN: Object Detection via Region-based Fully Convolutional Networks (NeurIPS 2016 · similarity 0.87)
- SCNN: A General Distribution Based Statistical Convolutional Neural Network with Application to Video Object Detection (AAAI 2019 · similarity 0.86)
- TGBFormer: Transformer-GraphFormer Blender Network for Video Object Detection (AAAI 2025 · 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.