动态对称块SAP迭代重建方法

SAP Iterative Reconstruction of Dynamic Symmetry Block

SAP(String-Averaging Projection)是图像重建的一类超平面投影迭代方法,将String顺序投影后加权平均进行迭代,以减小重建误差,提高成像质量.DSAP(Dynamic String-Averaging Projection)算法是在SAP算法基础上的改进,将String扩展为分块String,然后对分块String的投影值加权平均.将DSAP算法与块结构、射线间的对称关系相结合,构造出动态对称块SAP迭代方法,讨论了其几何意义和收敛性质,并给出了数据实验.实验结果表明:动态对称块SAP迭代方法与经典ART及对称块算法和块内SAP算法相比较,能更好地在空间分辨率和密度分辨率之间取得良好的折衷.

SAP is a class of hyperplane projection iterative method of image reconstruction.It takes a weighted average of the projection values obtained by sequential iteration.This iterative algorithm can reduce the error of reconstruction result and improve the quality of image.DSAP algorithm which is based on the SAP algorithm is further improved,and it takes the weighted average again after taking respectively a weighted average of projection values of block strings,which are the extension of the string.In this paper,the DSAP algorithm was combined block structure and symmetric relations between rays,the dynamic symmetric block SAP iteration method was constructed.It discussed geometric meaning and convergence property,and the experimental data were provided.The experimental results show that compared to the classical ART iterative algorithm,symmetric block iterative algorithm and SAP algorithm,the dynamic symmetric block SAP iteration method can better in a good compromise between the spatial resolution and density resolution.