Human Pose Manipulation for Synthetic Actors

Abstract

Human pose manipulation is the controlling of a person’s body in a static pic- ture and making him do whatever poses or actions that we force on him.This makes it possible bring back old movie actors to new movies or even any person we are able to reach to be inside the movie and move his body as we want.In this project , we present how to achieve pose manipulation using deep learning.The project contains more than one deep learning model to solve different problems first like pose estimation as we need poses to force the person to do.We used minimal resources like a low resolution camera and not much computational power to achieve the pose manipulation.