Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach

Many facial landmark methods based on convolutional neural networks (CNN) have been proposed to achieve favorable detection results. However, the instability landmarks that occur in video frames due to CNNs are extremely sensitive to input image noise. To solve this problem of landmark shaking, this...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Bing-Fei Wu, Bo-Rui Chen, Chun-Fei Hsu
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Facial landmark detection lightweight U-Net fast optical flow dynamic routing landmark stabilization Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/37766ec168a34fc7b23f0e230dc6c1d3
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:37766ec168a34fc7b23f0e230dc6c1d3
record_format	dspace
spelling	oai:doaj.org-article:37766ec168a34fc7b23f0e230dc6c1d32021-11-23T00:01:08ZDesign of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach2169-353610.1109/ACCESS.2021.3077479https://doaj.org/article/37766ec168a34fc7b23f0e230dc6c1d32021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9422692/https://doaj.org/toc/2169-3536Many facial landmark methods based on convolutional neural networks (CNN) have been proposed to achieve favorable detection results. However, the instability landmarks that occur in video frames due to CNNs are extremely sensitive to input image noise. To solve this problem of landmark shaking, this study proposes a simple and effective facial landmark detection method comprising a lightweight U-Net model and a dynamic optical flow (DOF). The DOF uses the fast optical flow to obtain the optical flow vector of the landmark and uses dynamic routing to improve landmark stabilization. A lightweight U-Net model is designed to predict facial landmarks with a smaller model size and less computational complexity. The predicted facial landmarks are further fed to the DOF approach to deal with the unstable shaking. Finally, a comparison of several common methods and the proposed detection method is made on several benchmark datasets. Experimental evaluations and analyses show that not only can the lightweight U-Net model achieve favorable landmark prediction but also the DOF stabilizing method can improve the robustness of landmark prediction in both static images and video frames. It should be emphasized that the proposed detection system exhibits better performance than others without requiring heavy computational loadings.Bing-Fei WuBo-Rui ChenChun-Fei HsuIEEEarticleFacial landmark detectionlightweight U-Netfast optical flowdynamic routinglandmark stabilizationElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 68737-68745 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Facial landmark detection lightweight U-Net fast optical flow dynamic routing landmark stabilization Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Facial landmark detection lightweight U-Net fast optical flow dynamic routing landmark stabilization Electrical engineering. Electronics. Nuclear engineering TK1-9971 Bing-Fei Wu Bo-Rui Chen Chun-Fei Hsu Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach
description	Many facial landmark methods based on convolutional neural networks (CNN) have been proposed to achieve favorable detection results. However, the instability landmarks that occur in video frames due to CNNs are extremely sensitive to input image noise. To solve this problem of landmark shaking, this study proposes a simple and effective facial landmark detection method comprising a lightweight U-Net model and a dynamic optical flow (DOF). The DOF uses the fast optical flow to obtain the optical flow vector of the landmark and uses dynamic routing to improve landmark stabilization. A lightweight U-Net model is designed to predict facial landmarks with a smaller model size and less computational complexity. The predicted facial landmarks are further fed to the DOF approach to deal with the unstable shaking. Finally, a comparison of several common methods and the proposed detection method is made on several benchmark datasets. Experimental evaluations and analyses show that not only can the lightweight U-Net model achieve favorable landmark prediction but also the DOF stabilizing method can improve the robustness of landmark prediction in both static images and video frames. It should be emphasized that the proposed detection system exhibits better performance than others without requiring heavy computational loadings.
format	article
author	Bing-Fei Wu Bo-Rui Chen Chun-Fei Hsu
author_facet	Bing-Fei Wu Bo-Rui Chen Chun-Fei Hsu
author_sort	Bing-Fei Wu
title	Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach
title_short	Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach
title_full	Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach
title_fullStr	Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach
title_full_unstemmed	Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach
title_sort	design of a facial landmark detection system using a dynamic optical flow approach
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/37766ec168a34fc7b23f0e230dc6c1d3
work_keys_str_mv	AT bingfeiwu designofafaciallandmarkdetectionsystemusingadynamicopticalflowapproach AT boruichen designofafaciallandmarkdetectionsystemusingadynamicopticalflowapproach AT chunfeihsu designofafaciallandmarkdetectionsystemusingadynamicopticalflowapproach
_version_	1718417368423596032

Design of a Facial Landmark Detection System Using a Dynamic Optical Flow Approach

Ejemplares similares