Three-dimensional virtual character intelligent skinning method

The invention belongs to the technical field of three-dimensional dynamic graph processing, and particularly relates to a three-dimensional virtual character intelligent skinning method based on deep learning, which particularly comprises the steps: skeleton generation (geometric contraction, edge folding and skeleton optimization) and skinning intelligent prediction (feature extraction, model construction, model training and skinning prediction). According to the method, skeleton generation, binding and skin weight assignment work can be automatically completed only through the grid model of the three-dimensional virtual character, so that a good effect can be achieved when the method is applied to grid models with different complexity degrees, skeleton binding and weight assignment are directly completed for the constructed three-dimensional virtual character, and the method is high in practicability and easy to popularize. When the method is applied to a high-precision three-dimensional model with a huge n