Attitude-guided virtual human body image generation method and system based on structural similarity and electronic equipment

Abstract

The invention discloses an attitude-guided virtual human body image generation method and system based on structural similarity and electronic equipment. The method comprises the following steps: step1, acquiring a source human body image and a target human body image; acquiring a target posture image according to the target human body image; step 2, inputting the source human body image and thetarget posture image in the step 1 into a pre-constructed convolutional neural network with an encoder-decoder structure to obtain a virtual target human body image; and step 3, constructing a loss function based on the virtual target human body image acquired in the step 2 and the target human body image acquired in the step 1, and performing iterative optimization on the pre-constructed convolutional neural network with an encoder-decoder structure. After a preset number of iterations is reached, the optimized convolutional neural network with an encoder-decoder structure is obtained, and the convolutional neural network is used for realizing virtual generation of a real scene human body image of a target posture. According to the invention, the real scene human body image of a more realtarget posture can be generated.

Description

technical field [0001] The invention belongs to the interdisciplinary technical field of computer vision and computer graphics, and in particular relates to a method, system and electronic equipment for generating a virtual human body image based on structural similarity guided by gestures. Background technique [0002] Pose-guided virtual human body image generation refers to a given human body image in a real scene and any target human body pose image, and under the guidance of the target human body pose image, a real scene human body image with the target human body pose is generated. This task has many meaningful real-world applications, such as: film production can use this method to directly manipulate the pose of characters; human body pose estimation generates rare but important human body image data. [0003] There are two problems in the current generation of virtual human body images: [0004] (1) Due to the large discrepancy between the source body image and the...

AI Technical Summary

Problems solved by technology

Aiming at the problem that the existing method ignores the special structural features of the human body in the process of generating the virtual human body image, and the generated target human body image lacks more detailed texture information, the invention proposes a method for generating a virtual human body image based on structural similarity, which can generate more realistic Real-world human images of target poses

Images