摘要
Real-world face detection and alignment demand an advanced discriminative model to address challenges by pose, lighting and expression. Recent studies have utilized the relation between face detection and alignment to make models computationally efficiency, but they ignore the connection between each cascade CNNs. In this paper, we combine detection, calibration and alignment in each cascade structure and propose an End-To-End cascade Online Hard Example Mining (OHEM) for training, which expert in accelerating convergence. Experiments on FDDB and AFLW demonstrate considerable improvement on accuracy and speed.
源语言 | 英语 |
---|---|
主期刊名 | Proceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017 |
出版商 | Institute of Electrical and Electronics Engineers Inc. |
页 | 35-40 |
页数 | 6 |
ISBN(电子版) | 9781538626368 |
DOI | |
出版状态 | 已出版 - 2 7月 2017 |
活动 | 7th International Conference on Virtual Reality and Visualization, ICVRV 2017 - Zhengzhou, 中国 期限: 21 10月 2017 → 22 10月 2017 |
出版系列
姓名 | Proceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017 |
---|
会议
会议 | 7th International Conference on Virtual Reality and Visualization, ICVRV 2017 |
---|---|
国家/地区 | 中国 |
市 | Zhengzhou |
时期 | 21/10/17 → 22/10/17 |
指纹
探究 'End-To-end cascade cnn for simultaneously face detection and alignment' 的科研主题。它们共同构成独一无二的指纹。引用此
Zhao, S., Song, H., Cong, W., Qi, Q., & Tian, H. (2017). End-To-end cascade cnn for simultaneously face detection and alignment. 在 Proceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017 (页码 35-40). 文章 8719201 (Proceedings - 2017 International Conference on Virtual Reality and Visualization, ICVRV 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICVRV.2017.00016