Initial evaluation of discrete orthogonal basis reconstruction of ECT images

Discrete orthogonal basis restoration (DOER) is a linear, non-iterative, and robust method for solving inverse problems for systems characterized by shift-variant transfer functions. This simulation study evaluates the feasibility of using DOER for reconstructing emission computed tomographic (ECT) images. The imaging system model uses typical SPECT parameters and incorporates the effects of attenuation, spatially-variant PSF, and Poisson noise in the projection process. Sample reconstructions and statistical error analyses for a class of digital phantoms compare the DOER performance for Hartley and Walsh basis functions. Test results confirm that DOER with either basis set produces images with good statistical properties. No problems were encountered with reconstruction instability. The flexibility of the DOER method and its consistent performance warrants further investigation of DOER as a means of ECT image reconstruction