A face key point tracking method and system for continuous frame video streams

Abstract

The present invention discloses a face key point tracking method and system for continuous frame video streams. The face picture is sent into a parameter model to obtain the original face key points; the original face key points are smoothed and the original coordinates are replaced. Points, get the preliminary stable face key points; track the movement direction and running distance of the face key points in the front and rear frames of face pictures, and perform abnormal point detection and correction on the preliminarily stable face key points. The present invention is simple to implement and has wide applicability, and can effectively ensure that the key points of the human face are displayed stably and in real time on the continuous video stream, thereby improving the human-computer interaction experience.

Description

technical field [0001] The present invention relates to the field of image processing, in particular to a method and system for tracking human face key points of continuous frame video streams. Background technique [0002] The key points of the face include the coordinate representation of key parts such as eyes, eyebrows, nose, mouth and facial contour on the face picture. Face key point tracking technology is applied to various expression transmission scenarios under human-computer interaction, which is of great significance to the research of digital images and human-computer interaction fields. For example, virtual expression recording and real-time virtual character interaction rely on face key point tracking technology . [0003] The face key points are a set of coordinate points obtained by analyzing and calculating the face image by the parametric model, and a series of face key points obtained in the streaming video frame are easily affected by the parametric mode...

AI Technical Summary

Problems solved by technology

[0004] The present invention proposes a face key point tracking method and system for continuous frame video streams, which solves the problem that there may be uncertain deviations in the face key points calculated and analyzed by the parameter model in the prior art, and directly obtains the face key points When the virtual expression is driven, the virtual expression will show a shaking effect visually, resulting in a poor human-computer interaction experience.

Images