Image Segmentation Based on Shape Space Modeling

A new image segmentation method based on the active contour is presented in this paper. If we define a shape space as a set of all possible variations from an initial curve, a shape matrix represents variations of an initial curve with a few parameters and spans a subspace of the shape space. If we assume that the universal shape space is linear, it can be decomposed into the column space and the left null space of the shape matrix. Therefore, we perform image segmentation in two subspaces separately. In the proposed method, a shape space vector in the column space describes changes from an initial curve to the imaginary feature curve and a dynamic graph search algorithm describes the detailed shape of an object in the left null space. Since we employ the shape matrix, the proposed algorithm can ignore unwanted feature points generated by low-level image processing.