Method for generating model of sculpture of face, computing device, and non-transitory storage medium

A facial sculpture model generating device comprising:

• at least one processor;
• and
• a storage device storing one or more programs which when executed by the at least one processor, causes the at least one processor to:
• train a predetermined generative adversarial network (GAN) based on a three-dimensional (3D) face dataset to obtain an initial sculpture generation model, wherein the 3D face dataset comprises 3D face data of multiple 3D face images;
• perform a curvature conversion on each of the multiple 3D face images to obtain a distribution map of curvature value;
• add the distribution map of curvature value of each of the multiple 3D face images as attention information to the initial sculpture generation model, to train and generate a facial sculpture generation model based on curvature value attention mechanism;
• add a blurring operator of a predetermined value to the facial sculpture generation model for training to smooth face curvature values of a facial sculpture model generated by the facial sculpture generation model;
• and
• receive a target 3D face data and predetermined face curvature parameters, and input the target 3D face data and the predetermined face curvature parameters into the facial sculpture generation model to generate a corresponding facial sculpture model.

A non-transitory storage medium having stored thereon instructions that, when executed by a processor of a computing device, causes the computing device to perform a facial sculpture model generating method, the method comprising:

• training a predetermined generative adversarial network (GAN) based on a three-dimensional (3D) face dataset to obtain an initial sculpture generation model, wherein the 3D face dataset comprises 3D face data of multiple 3D face images;
• performing a curvature conversion on each of the multiple 3D face images to obtain a distribution map of curvature value;
• adding the distribution map of curvature value of each of the multiple 3D face images as attention information to the initial sculpture generation model, to train and generate a facial sculpture generation model based on curvature value attention mechanism;
• adding a blurring operator of a predetermined value to the facial sculpture generation model for training to smooth face curvature values of a facial sculpture model generated by the facial sculpture generation model;
• and
• receiving a target 3D face data and predetermined face curvature parameters, and inputting the target 3D face data and the predetermined face curvature parameters into the facial sculpture generation model to generate a corresponding facial sculpture model.