Game character action recognition and generation method based on neural network

Abstract

The invention discloses a game character action recognition and generation method based on a neural network, and the method comprises the following steps: collecting a human motion video, and segmenting the human motion video into continuous frames of images according to a frame rate; adopting a human body key point neural network to detect and extract human body key point 2D coordinate data in the image of any sequence frame one by one; inputting the 2D coordinate data of the human body key points of any sequence frame into a posture prediction neural network to obtain corresponding 3D posture data; and binding the 3D posture data to the animation skeleton to obtain game character action data. By means of the scheme, the method has the advantages of being simple in logic, accurate in recognition, high in recognition effect, small in calculation workload and the like, and has high practical value and promotional value in the technical field of computer vision.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a neural network-based game character action recognition and generation method. Background technique [0002] In recent years, with the development and application of computer technology and AI artificial intelligence, human motion reconstruction technology based on computer vision has risen rapidly and received extensive attention. At present, human motion reconstruction based on computer vision is still a very challenging topic in computer vision, which involves image processing, pattern recognition, artificial intelligence and other disciplines, and has extensive experience in digital film, animation production, game development and other fields. Application prospects. [0003] This article is mainly applied to the production of game actions. Among them, the traditional game action production techniques mainly include the following three types: [0004] The first is ...

AI Technical Summary

Problems solved by technology

[0006] The third method is manual production by professionals; this method requires professional animators to manually produce corresponding movements frame by frame, which takes a long time and is inefficient, and high-quality movements require high production staff

There are some approaches to predict 3D pose from a given set of 2D keypoints by exploiting image features and 2D ground truth, or by simply predicting their depth; traditional video pose estimation mostly works on single-frame data, which leads to The data between frames is independent and discontinuous

Images