A team from Peking University has developed a deep learning model for age estimation using non-registered 3D face point clouds. They trained the model on over 16,000 instances of 3D face point cloud data and used coordinate-wise monotonic transformations to preserve age-related features while distorting faces. The algorithm can isolate age-related facial features and achieve accurate age estimation with an average absolute error of 2.5 years. The researchers highlighted the importance of protecting facial data due to privacy concerns and proposed a facial data protection guideline to manage facial data centers or public datasets. The study aims to broaden public access to face datasets while minimizing privacy risks. This research represents a novel approach to applying deep learning directly to 3D face point cloud data for face recognition and facial expression detection. The findings have implications for enhancing data security and privacy in the field of biometrics and deep learning.
Source link
Source link: https://www.biometricupdate.com/202405/chinese-researchers-test-point-cloud-based-facial-age-estimation-process
Chinese researchers develop facial age estimation using point cloud #innovation
![Chinese researchers test point cloud-based facial age estimation process](https://i0.wp.com/webappia.com/wp-content/uploads/2024/05/3D-face-point-cloud.png?fit=758%2C530&quality=80&ssl=1)
GIPHY App Key not set. Please check settings