FEATURE AND MODEL MUTUAL MATCHING FACE TRACKING METHOD BASED ON INCREMENT PRINCIPAL COMPONENT ANALYSIS

摘要

Disclosed is a feature and model mutual matching face tracking method based on on-line increment principal component analysis. The method comprises the following steps: performing off-line modeling on multiple face images to obtain a model matching (CLM) model A; performing key point detection on each frame of a face video to be tracked, and combining a set of all key points and robust descriptors thereof into a key point model B; performing, on the basis of the key point model B, key point matching on each frame of the face video to be tracked to obtain an initial face gesture parameter set in each frame of face image; performing, by using the model A, CLM face tracking on the face video to be tracked; performing re-tracking according to the initial face gesture parameter set and an initial tracking result; and updating the model A, and repeating the steps to obtain a final face tracking result. The present invention solves the problem of tracking losing occurred when a variation between adjacent frames in a target image is large during CLM face tracking, thereby improving the tracking accuracy.