Image Reconstruction for Pet Using Fuzzy Potential

Mondal, Partha Pratim and Rajan, K (2004) Image Reconstruction for Pet Using Fuzzy Potential. In: 14th IEEE Signal Processing Society Workshop on Machine Learning for Signal Processing, 29 September - 1 October 2004, Sao Luis, Brazil, pp. 433-442.

Preview

PDF Image_Feb16th.pdf Download (392kB)

Abstract

Maximum likelihood (ML) and Maximum a-posteriori (MAP) are the most widely used algorithms for emission tomography (ET). MAP-approach heavily depends on Gibbs hyperparameter for noise free reconstruction. Choosing a correct hyperparameter is difficult and time consuming task. Recently proposed median root prior (MRP) algorithm is a good alternative, but are prone to step like streaking effect. In this research work, a fuzzy logic based approach is proposed to overcome these shortcomings. Unlike traditional potential function, a fuzzy potential is used for modeling inter pixel interaction. Two basic operations viz. edge detection and fuzzy smoothing are performed sequentially during each iteration. The first operation is employed for the detection of edges (if present) in all the eight directions of a 3 x 3 neighborhood window. The second operation uses this edge information to perform fuzzy smoothing. Due to recursive nature of the reconstruction algorithm, both these operations are employed iteratively to reduce heavy noise produced due to dimensional instability [10]. Simulated experimental results are obtained to show the feasibility of the proposed approach. These algorithms are also compared with other approaches such as, MAP and MRP by numerical measures and visual inspection. Algorithm evaluation shows promising results.

Item Type:	Conference Paper
Publisher:	IEEE
Additional Information:	copyright 2007 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Department/Centre:	Division of Physical & Mathematical Sciences > Physics
Date Deposited:	22 Aug 2008
Last Modified:	19 Sep 2010 04:35
URI:	http://eprints.iisc.ac.in/id/eprint/9873

Actions (login required)

View Item