Research map / Detection and segmentation / CNN features
MogFace: Towards a Deeper Appreciation on Face 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 |
|---|---|---|---|
| Add Ali-AMS label assignment (ablation) | SFD with ResNet50 · Wider Face val Hard | AP: 86.5 -> 87.3 | Helped |
| Use HAMBox instead of standard label assignment | SFD with ResNet50 · Wider Face val Hard | AP: 86.5 -> 86.9 | Helped |
| Use ATSS instead of standard label assignment | SFD with ResNet50 · Wider Face val Hard | AP: 86.5 -> 86.3 | Hurt |
| Use OTA instead of standard label assignment | SFD with ResNet50 · Wider Face val Hard | AP: 86.5 -> 84.7 | Hurt |
| Add SSE to control ground-truth scale distribution for pyramid layers (ablation) | SFD with ResNet50 · Wider Face val Easy | AP: 94.6 -> 95.6 | Helped |
| Add SSE to control ground-truth scale distribution for pyramid layers (ablation) | SFD with ResNet50 · Wider Face val Medium | AP: 93.4 -> 94.1 | Helped |
| Use multi-scale training rather than the baseline augmentation | SFD with ResNet50 · Wider Face val Hard | AP: 81.4 -> 83.6 | Helped |
| Use data-anchor-sampling rather than the baseline augmentation | SFD with ResNet50 · Wider Face val Hard | AP: 81.4 -> 86.5 | Helped |
| Add HCAM with neighbour context sizes 3 and 5 (ablation) | SFD with ResNet50 · Wider Face val Hard | AP: 86.5 -> 87.4 | Helped |
| Add HCAM with neighbour context sizes 3 and 5 (ablation) | SFD with ResNet50 · Wider Face val Hard | number of false alarms: 948 -> 192 | Helped |
| Use only context size 3 instead of sizes 3 and 5 (ablation) | HCAM · Wider Face val Hard | AP: 87.4 -> 86.8 | Hurt |
| Use only context size 5 instead of sizes 3 and 5 (ablation) | HCAM · Wider Face val Hard | AP: 87.4 -> 87.0 | Hurt |
Closest papers by meaning
- Sample and Computation Redistribution for Efficient Face Detection (ICLR 2022 · similarity 0.89)
- Finding Tiny Faces (CVPR 2017 · similarity 0.88)
- WIDER FACE: A Face Detection Benchmark (CVPR 2016 · similarity 0.88)
- Group Sampling for Scale Invariant Face Detection (CVPR 2019 · similarity 0.87)
- KPNet: Towards Minimal Face Detector (AAAI 2020 · similarity 0.87)
- ContextShift: A Controlled Benchmark for Context Dependence in Object Detection (NeurIPS 2026 · similarity 0.84)
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.