This is a research course that covers topics of interest in computer vision, including image denoising/deblurring, image segmentation/object detection, and image registration / matching. The emphasis will be on creating mathematical models via the framework of Bayesian estimation theory, analyzing these models, and constructing computational algorithms to realize these models. Techniques from calculus of variations, differential geometry, and partial differential equations will be built up as the need arises. AMCS 396 was previously entitled "Math Models in Computer Vision and Image" during Spring 2012.