Subjective Surfaces: A Geometric Model for Boundary Completion

摘要: We present a geometric model and computational method for segmentation of images with missing boundaries. In many situations, the human visual system fills in gaps edges boundaries, building completing information that is not present. Boundary completion presents considerable challenge computer vision, since most algorithms attempt to exploit existing data. A large body work concerns models, which postulate how construct datas these models are often trained specific particular images. this paper, we take following, alternative perspective: consider given reference point within image, then develop an algorithm tries build on basis view available as boundary data algorithm. Starting from view, surface constructed. It evolved mean curvature flow metric induced by image until piecewise constant solution reached. test modal completion, amodal texture segmentation. extend 3D order extract shapes low signal/noise ratio ultrasound volumes. Results echocardiography fetal echography also presented.