A new approach to medical image reconstruction

The authors propose a multiobjective optimization method using the weighted sum problem in order to deal with image reconstruction from projections. In their scheme, after determining the objective functions, a satisfactory solution of the decision maker can be derived by updating the weighted coefficients simultaneously with the tradeoff between the objective functions. Furthermore, the noninferiority of the generated solution in each iteration is guaranteed. As has been shown, the multiobjective optimization-based image reconstruction approach is superior to the least-squares and convolution methods. The authors can say that the multiobjective approach makes traditional image reconstruction from projections flexible and robust to uncertainty and multiplicity of reconstruction objectives. They are convinced that the proposed method and its extension will become an efficient tool for design of new CT scanners.