Research Article
FLASH: Facial Landmark Detection Using Active Shape Model and Heatmap Regression
@INPROCEEDINGS{10.1007/978-3-031-47359-3_13, author={Nguyen Van Nam and Ngo Thi Ngoc Quyen}, title={FLASH: Facial Landmark Detection Using Active Shape Model and Heatmap Regression}, proceedings={Industrial Networks and Intelligent Systems. 9th EAI International Conference, INISCOM 2023, Ho Chi Minh City, Vietnam, August 2-3, 2023, Proceedings}, proceedings_a={INISCOM}, year={2023}, month={10}, keywords={facial landmarks heatmap regression shape fitting coordination regression}, doi={10.1007/978-3-031-47359-3_13} }- Nguyen Van Nam
Ngo Thi Ngoc Quyen
Year: 2023
FLASH: Facial Landmark Detection Using Active Shape Model and Heatmap Regression
INISCOM
Springer
DOI: 10.1007/978-3-031-47359-3_13
Abstract
Detection of facial landmarks is a critical task for human face identification, emotion recognition in autopilot and real-time visual monitoring applications. This is really challenging due to the high number of discrete landmarks spreading over the face which is of different shapes and may be occluded or obscured. Many methods have been proposed over the years including ASMNet and AnchorFace. However, their performance is still limited in terms of both accuracy and efficiency. In this paper, we propose a novel method for facial landmark detection based on active shape model and heatmap called FLASH. The heatmap aims to highlight the important landmarks. Meanwhile, the shape model helps to conform the distribution of such landmarks. FLASH has been evaluated on two public datasets 300W-Challenging, WFLW and achieved a normalized mean square error (NME) of 6.67%, 7.34% correspondingly, which outperforms most existing methods. Specifically, this is much better than the recent ASMNet method with a NME of 8.20%, 10.77% on the two datasets, respectively. This is also comparable to the state of the art AnchorFace with a NME of 6.19%, 4.62%, correspondingly. The source code of FLASH is also publicly available.
