Research map / Detection and segmentation / CNN features
M2Det: A Single-Shot Object Detector Based on Multi-Level Feature Pyramid Network
Read the paperCheck 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 |
|---|---|---|---|
| Add deconvolution layers (ablation) | SSD baseline · COCO minival | AP: 25.8 -> 27.5 | Helped |
| Use 8 s-TUMs (ablation) | SSD-based detector · COCO minival | AP: 27.5 -> 30.6 | Helped |
| Replace s-TUMs with TUMs (ablation) | M2Det variant · COCO minival | AP: 30.6 -> 30.8 | Helped |
| Feed the base feature to each TUM (ablation) | M2Det variant · COCO minival | AP: 30.8 -> 32.7 | Helped |
| Add SFAM (ablation) | M2Det variant · COCO minival | AP: 32.7 -> 33.2 | Helped |
| Replace VGG-16 with ResNet-101 (ablation) | M2Det · COCO minival | AP: 33.2 -> 34.1 | Helped |
| Increase channels from 256 to 512 with 2 TUMs (ablation) | M2Det MLFPN · COCO minival | AP: 30.5 -> 32.1 | Helped |
| Increase channels from 256 to 512 with 4 TUMs (ablation) | M2Det MLFPN · COCO minival | AP: 31.8 -> 33.4 | Helped |
| Increase channels from 256 to 512 with 8 TUMs (ablation) | M2Det MLFPN · COCO minival | AP: 33.2 -> 34.0 | Helped |
| Increase TUM count from 2 to 4 at 256 channels (ablation) | M2Det MLFPN · COCO minival | AP: 30.5 -> 31.8 | Helped |
| Increase TUM count from 4 to 8 at 256 channels (ablation) | M2Det MLFPN · COCO minival | AP: 31.8 -> 33.2 | Helped |
| Increase TUM count from 8 to 16 at 256 channels (ablation) | M2Det MLFPN · COCO minival | AP: 33.2 -> 33.6 | Helped |
Closest papers by meaning
- Feature Pyramid Networks for Object Detection (CVPR 2017 · similarity 0.93)
- AugFPN: Improving Multi-Scale Feature Learning for Object Detection (CVPR 2020 · similarity 0.91)
- EfficientDet: Scalable and Efficient Object Detection (CVPR 2020 · similarity 0.90)
- You Only Look One-Level Feature (CVPR 2021 · similarity 0.90)
- Fine-Grained Dynamic Head for Object Detection (NeurIPS 2020 · similarity 0.89)
- Scale-Equalizing Pyramid Convolution for Object Detection (CVPR 2020 · 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.