An effective reconstruction algorithm of attenuation map from post-injection PET transmission scans

We present an effective reconstruction algorithm of attenuation map from post-injection transmission scans. The post-injection transmission scan data are modeled as a collection of independent Poisson observations which consist of true transmission, cross-containment emission, and random coincidences. The mean values of cross-containment emission and random data are estimated by using Luk's approximation and Zaers' correction method, respectively. The reconstruction of the unknown attenuation map is also modeled using a Markov random field. A pre-conditioned conjugate gradient algorithm is then used to compute a maximum a posteriori estimate of the attenuation map by maximizing over the posterior density. Evaluations are conducted on clinical data collected from a CTI EXACT HR+ scanner. Experimental results show that the reconstructed results of attenuation map by the described reconstruction algorithm match well with theoretical values of soft tissue, bone and lung under 512 KeV photons.