Pose Estimation and 3D Model Overlay in Real Time for Applications in Augmented Reality

Abstract

AbstractPose estimation in real time is a challenging research area with numerous applications. Obtaining detected key points in each pose presents us with useful applications, one such application being in the augmented reality domain. This paper presents an approach to 3D model overlay on an image with pose key point information. The model used in this work is based on the paper titled Multi-Person Pose Estimation. An FCN is employed for extracting key point information from the data. The method used in this work involves intersecting the image plane and the desired 3D model in real time and orienting the model correctly in accordance with the key point information. The performance will greatly depend on the GPU specifications.KeywordsAugmented realityPose estimation3D modelReal-time applications