Supplemental Material to Accompany “ Human Ear Recognition in 3 D ”

A. Detection 1) Edge clustering result: Three examples of clustering results are shown in the first row of Figure 1, in which each cluster is bounded by a rectangular box and each edge segment is shown in different color. The second row shows the extracted regions-of-interest bounded by boxes overlaid on the color images. From Figure 1, we observe that the ear region is correctly identified. 2) Application of the global-to-local registration procedure with the optimization formulation: The proposed global-to-local procedure with the optimization formulation can also be applied to handle the non-rigid shape registration. We performed experiments on two simple shapes of the hand and dude 1.