An Automated Generation from Video to 3D Character Animation using Artificial Intelligence and Pose Estimate

Abstract

This paper presents a novel approach to automatically generate 3D character animation from video using artificial intelligence and pose estimation [3]. The proposed system first extracts the pose information fromthe input videousing a pose estimation model [2]. Then, an artificial neural network is trained to generate the corresponding3Dcharacter animation based on the extracted pose information [1]. The generated animation is then refined usingaset of animation filters to enhance the quality of the final output. Our experimental results demonstrate theef ectiveness of the proposed approach in generating realistic and natural-looking 3D character animations fromvideo input [4]. This automated process has the potential to greatly reduce the time and ef ort required for creating3D character animations, making it a valuable tool for the entertainment and gaming industries.