Efficient reconstruction of moment-coded noisy medical images using maximum entropy
D.A. Karras · 2006
The information-theoretic method of maximum entropy (ME) is applied in a novel way to the problem of reconstructing noisy endoscopic grey-scale images from a finite set of its geometric moments (GMs). Simulation results, using representative medical images from endoscopic video sequences acquired during colonoscopy examinations, demonstrate the superiority of the ME method, with almost threefold compression ratios achieved, over the Legendre moments (LMs) approach.