Three-dimensional head processing method and training method and device of three-dimensional head processing model

• Main Authors: YE ZHILING, ZHANG LIANGGUO
• Format: Patent
• Language: Chinese; English
• Published: 30.04.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The embodiment of the invention provides a three-dimensional head processing method and a training method and device of a three-dimensional head processing model. The processing method comprises the following steps: acquiring a first driving voice to be processed and a first camera viewpoint; generating a first high-dimensional spatial feature of the first driving voice; generating first voxel information of M first target voxels according to the first high-dimensional spatial features and the first camera viewpoint; the first target voxel is a voxel corresponding to a first camera viewpoint and a target imaging plane, the target imaging plane is an imaging plane of a first target three-dimensional head to be driven, and M is an integer greater than 1; and performing voxel rendering processing on the target imaging plane according to the first voxel information to obtain a first target three-dimensional head. According to the embodiment of the invention, the driving effect of the three-dimensional head is imp