요약
Abstract
Ⅰ. Introduction
Ⅱ. Related Works
2.1 State-of-the-arts face detection
2.2 Lightweight face detection
Ⅲ. EdgeFace
3.1 EdgeFaceNet
3.2 Lightweight Feature Pyramid Network
3.3 Efficient Context Module
3.4 Multi-task Learning
Ⅳ. Experiments
4.1 Implementation Details
4.2 Ablation Study
4.3 Comparison with Other Frameworks
4.4 Speed on Edge Devices
Ⅴ. Conclusion
REFERENCES