A Fast algorithm of PET System Response Matrix based on Straight Line Truncation

TL;DR

This paper introduces a Straight Line Truncation (SLT) method that significantly accelerates PET System Response Matrix calculations and reduces storage requirements, enabling more efficient iterative image reconstruction.

Contribution

The paper presents a novel SLT algorithm that improves SRM computation speed by up to 23 times and reduces storage space to one sixty-fourth of traditional methods.

Findings

SLT algorithm is six times faster than traditional Ray-Tracing.

SLT reduces storage space to 1/64 of traditional Ray-Tracing.

SLT achieves 23 times speedup at small angles.

Abstract

In the calculation of Positron Emission Tomography (PET) image reconstruction, System Response Matrix (SRM) characterizes the numerical relationship between measurement space and image space. Due to a significant amount of calculation and low efficiency of the current SRM algorithm, the application of iterative reconstruction in the medical and industrial fields is limited. The time and space of computing SRM are the most important factors that researchers are concerned about. In this paper, the Straight Line Truncation (hereinafter referred to as SLT) method is applied to calculate the SRM. We verified that it improves the speed of calculating the SRM and reduces its spatial complexity. The experimental results show that the forward projection speed of SLT algorithm is six times more than that of the traditional Ray-Tracing algorithm. Besides, the forward projection speed of the small angle of SLT is about 23 times that of the traditional Ray-Tracing algorithm, and the storage space of SLT is about one sixty-fourth of the storage space of traditional Ray-Tracing algorithm.