A fast first-order continuation total variation algorithm for image denoising

Dongfang Chen, Ningtao Zhang · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2013

This paper describes a new denoising algorithm which named Continuation total-variation denoising (CTVD) algorithm. The algorithm is based on the novel Nesterov’s first-order method, most notably Nesterov’s smoothing technique, tailored to the image processing application. The algorithm also introduces a fixed point continuation technique to increase the convergence rate, and in particular when to deal with high dynamic range signals and largescale problems.

Read the paper · More papers on PaperTik