Modularized generation system for three-dimensional facial animation

The invention discloses a three-dimensional facial animation modular generation system, and relates to the technical field of three-dimensional facial animation generation, and the system comprises the steps: firstly carrying out the feature extraction, coding and quantification of a target audio stream through an audio feature encoder, and obtaining audio context feature information and emotion features; secondly, performing alignment synchronization on the audio context feature information and preset facial feature information by utilizing a facial animation decoder according to the emotional features, a facial node priori graph constructed based on large language model semantic knowledge and a style embedding vector constructed based on One-Hot coding to obtain facial node feature information aligned with the audio; and finally, performing independent reconstruction on the upper part and the lower part of the face by using a face region separation module according to the face node feature information align