Bin-Picking based on Harmonic Shape Contexts and Graph-Based Matching

In this work we address the general bin-picking problem where 3D data is available. We apply harmonic shape contexts (HSC) features since these are invariant to translation, scale, and 3D rotation. Each object is divided into a number of sub-models each represented by a number of HSC features. These are compared with HSC features extracted in the current data using a graph-based scheme. Results show that the approach is somewhat sensitive to noise, but works in presence of occlusion