Manifold valued image segmentation

Abstract

Image segmentation is the process of partitioning a image into different regions or groups based on some characteristics like color, texture, motion or shape etc. Segmentation is an intermediate process for a large number of applications including object recognition and detection. Active contour is a popular variational model for object segmentation in images, in which the user initializes a contour which evolves in order to optimize an objective function designed such that the desired object boundary is the optimal solution. Recently, imaging modalities that produce Manifold valued images have come up, for example, DT-MRI images, vector fields. The traditional active contour model does not work on such images.

In the work presented here we generalize the active contour model to work on Manifold valued images. Since usual gray-scale images are just an specific example of Manifold valued images, our method produce expected results on gray-scale images. As an application of proposed active contour model we we perform texture segmentation on gray-scale images by first creating an appropriate Manifold valued image. We demonstrate segmentation results for manifold valued images and texture images.

Diversity of the texture segmentation problem Inspired us to propose a new active contour model for texture segmentation where we find the background/foreground texture regions in a given image by maximizing the geodesic distance between the interior and exterior covariance matrices. We also provide results using proposed method.