Physics-Based Object Pose and Shape Estimation from Multiple Views

This paper presents a new algorithm for object pose and shape estimation from multiple views. Using a qualitative shape recovery scheme we first segment the image into parts which belong to a vocabulary of primitives. Based on the additional constraints provided by the qualitative shapes we extend our physics-based framework to allow object pose and shape estimation from stereo images where the two cameras have arbitrary relative orientations. We then generalize our algorithm to integrate measurements from multiple views. To recover more complex objects we generalize the definition for the global bending deformation. We also present an algorithm for model discretization which evenly tessellates the model surface. We demonstrate the usefulness of our technique in experiments involving real images from of a variety of object shapes which may be partially occluded.