Range Image Segmentation Using Principal Curvatures and Principal Directions

This paper proposes a novel range image segmentation method. Taking the peak area of the surface as the seed, we apply the region growing to exactly locate the edge according to the characteristics of the principal curvatures and its directions. We apply the edge detection technology to the region growth rules, so the boundary localization is precise. To obtain more stable edge information, a multi-scale fusion approach is proposed. The experiments results on footprint range images show that the segmented patches contain footprint biometric information well and set a reliable basis for the further recognition.