Total Variation Based Multiscale Representation
Leopold Matamba Messi (Mathematical Biosciences Institute, The Ohio State University)
(April 17, 2014 10:20 AM - 11:15 AM)
New imaging modalities are continuously developed in basic science research labs across the nation. In spite of the advances in imaging technologies, the signals that are recorded are still plagued with noise. High quality image denoising and segmentation algorithms are a necessity in every basic science lab that uses imaging as it primary tool of investigation. In this talk, I will review total variation minimization based methods for image denoising and segmentation. I will then discuss how one may use the ROF model to adaptively resolved the features in general images across scales, and conclude with an analysis of the convergence rate of the resulting multiscale decomposition scheme.