Multi-view face synthesis method and system based on generative adversarial network

The invention relates to a computer vision and pattern recognition technology, and discloses a multi-view face synthesis method and system based on a generative adversarial network. The method comprises the following steps: data preparation: acquiring data sets of different identities and postures; generating a front face, inputting a face image x, and generating a front face image x front face through a face normalization generator; front face editing: receiving the front face image x front face by a face editor, and controlling the front face image x front face by the face editor through a remote control module to generate a face image x rotation with a set angle; carrying out face comparison and recognition, wherein the face recognizer compares and recognizes the input face image x and the characteristics of x rotation which is generated through face editing and has the same deflection angle as x; updating parameters, namely updating the parameters according to a distributed training strategy; and iterativ