3D facial recognition in video conferencing based on Kinect depth images and covariance matrix method

Video quality and resolution have a significant impact on the accuracy of 3D facial recognition. Due to potential issues such as network latency, transmission loss, and image compression in video conferencing, the quality of different devices and cameras can also lead to differences in video quality. How to handle low-quality videos, improve image resolution, and reduce noise interference is an important challenge. In order to improve user experience, enhance conference security, and adapt to diverse scene requirements, a video conference 3D face recognition method based on Kinect depth images and covariance matrix method is proposed. Using Kinect depth imaging technology, collect 3D facial data from video conferences, and use Holt dual exponential smoothing algorithm to smooth and process the collected 3D facial data from video conferences. Using covariance matrix method to extract 3D facial features from video conferencing, applying support vector machine combined with deep learning models to achieve 3D facial recognition in video conferencing. The experimental results show that the proposed method has a good effect, and can effectively improve the accuracy and efficiency of 3D face recognition in video conferencing.