单光子发射断层成像的奇异值分解重建算法

Singular value decomposition reconstruction algorithm for single photon emission computed tomography

为了抑制单光子发射断层成像(SPECT)中噪声的影响,提高重建图像的质量和定量精度,应用奇异值分解(SVD)方法进行图像重建。对人体胸腔模型进行Monte-Carlo模拟计算,生成三维SPECT系统传输矩阵和模拟投影图像,求解系统传输矩阵的广义逆矩阵。在有噪声情况下,存在最佳保留奇异值数目,使重建图像质量达到最优。最佳保留奇异值数目的不同体现了噪声的差异。与常规重建方法进行比较,SVD重建算法具有更好的噪声抑制和重建图像质量,是一种值得关注的SPECT图像重建算法。

Singular value decomposition (SVD) was used for image reconstruction to reduce the influence of noise in single photon emission computed tomography (SPECT). The 3D system transition matrix and the projection data were produced by a Monte-Carlo simulation based on a human torso phantom. The optimized reserve singular value, which represents the noise level, gives the best reconstructed image quality when noise is present. Compared with conventional reconstruction methods, the SVD reconstruction algorithm more effectively reduces the influence of noise to greatly improve the reconstructed result.