Continuous Mixture Models for Feature Preserving Smoothing and Segmentation

Image smoothing and segmentation are fundamental tasks in Computer Vision and there are numerous algorithms that have been developed and applied to these tasks in various application domains. Several challenges remain unconquered. In this paper, we consider the challenge of achieving smoothing and segmentation while preserving complicated and detailed features present in the image, be it a gray level or a textured image. We present a novel approach that does not make use of any prior information about the objects in the image being processed. The key idea here is to model the derived local orientation information via a continuous mixture of appropriate basis functions. We present two such models; one involving a continuous mixture over the covariance matrices of Gaussian basis functions, and another involving a continuous mixture over the mean direction vectors of antipodally symmetric Watson basis functions. These continuous mixture models are then used to construct spatially varying kernels which are convolved with the input function to achieve feature preserving smoothing or segmentation as desired. We present numerous experimental results on real images, and compare and validate the proposed models on images drawn from Berkeley Segmentation Data Set quantitatively. Superior performance of our technique is depicted via comparison to the state-of-the-art algorithms in literature.