METHOD FOR EDITING FACIAL EXPRESSION BASED ON SINGLE CAMERA AND MOTION CAPTURE DATA

摘要

Disclosed is a method for editing a virtual facial expression, which comprises an offline process and an online process. The offline process comprises: step 1, utilizing a full face photo to generate a virtual three-dimensional face model of a user; step 2, decoupling motion capture data and separating a gesture from an expression; and step 3, constructing a facial subspace association model. The online process comprises: step 4, performing video analysis on a video image captured in a single camera, using an active appearance model to trace a three-dimensional space position and an orientation of a head and main facial features in a video, and automatically converting the features to two parts of control parameters; and step 5, using the motion capture data after decoupling to perform dynamic data filtering on expression control parameters with loud noise and low resolution, inputting signals after filtering into the facial subspace association model, computing and obtaining a global expression, and finally assigning the global expression to the virtual three-dimensional face generated in the offline process to implement editing of the virtual facial expression.