Abstract
We discuss a multiobjective image reconstruction from projections. The multiobjective decision model involves the entropy function of an image, the squared error function between the original projection data and the reprojection data due to the reconstructed image, and the peakedness function of the image. We develop a weighted-sum optimization algorithm for image reconstruction under these three conflicting objectives. The decision process is based on a weighted-sum optimization and a nonlinear programming algorithm. Comparisons of the multiobjective method with the convolution backprojection and simultaneous-algebraic-reconstruction technique algorithms are carried out with the use of computer-generated noise-free and noisy projections. An improvement in image reconstruction with the multiobjective method is demonstrated.
© 1991 Optical Society of America
Full Article | PDF ArticleMore Like This
Yuanmei Wang
J. Opt. Soc. Am. A 18(5) 1027-1032 (2001)
Hiroyuki Kudo and Tsuneo Saito
J. Opt. Soc. Am. A 8(7) 1148-1160 (1991)
Robert T. Smith, Csaba K. Zoltani, George J. Klem, and Monte W. Coleman
Appl. Opt. 30(5) 573-582 (1991)