Real-time Upper Body 3D Pose Estimation from a Single Uncalibrated Camera

Abstract

In this paper we introduce a real-time steering controller ensuring the reach of a (possible mobile) target position and orientation, without requiring to build/update the full trajectory to that target. We name it the funnelling control. The final orientation is achieved through the continuous adjustment of the heading direction. This control mode is proactive in the sense that it anticipates the success/failure of the reach and adjusts the desired speed accordingly. Both features rely on an heterogeneously sampled table of radial-tangential seek angles obtained when the controller reaches a desired position target without prescribed orientation. By construction, the control update has a constant computing cost, even with variable target characteristics. Its low update cost makes it particularly suited for controlling a large number of mobile entities in real-time. The present exposition is made for an obstacle-free context.This paper outlines a method of estimating the 3D pose of the upper human body from a single uncalibrated camera. The objective application lies in 3D Human Computer Interaction where hand depth information offers extended functionality when interacting with a 3D virtual environment, but it is equally suitable to animation and motion capture. A database of 3D body configurations is built from a variety of human movements using motion capture data. A hierarchical structure consisting of three subsidiary databases, namely the frontal-view Hand Position (top-level), Silhouette and Edge Map Databases, are pre-extracted from the 3D body configuration database. Using this hierarchy, subsets of the subsidiary databases are then matched to the subject in real-time. The examples of the subsidiary databases that yield the highest matching score are used to extract the corresponding 3D configuration from the motion capture data, thereby estimating the upper body 3D pose.