Development of an imaging algorithm for myocardial SPECT with a semiconductor detector

本研究的目的是开发一种新的基于半导体探测器的心肌单光子发射CT数据采集方案。一个带有半导体探测器的伽马相机使我们能够测量任意角度的投影数据，因为带有这个探测器的伽马相机比带有闪烁探测器的普通伽马相机小。提出的数据采集方案是在投影数据信噪比较高的角度测量投影数据。我们假设我们可以用25个投影来测量心脏。探测器在x-y平面上的位置分别为φ=-30、0、30、60和90度，与z轴的位置分别为0=30、60、90、120和150度。然后，我们确定了可接受的图像质量所需的最小投影数，并选择了最佳投影角度。选择投影角度的标准是每个投影的功率谱的积分值。我们根据这个积分值将投影数据按降序排列，得到4个投影子集。一个子集由三个投影组成，即积分值最大的投影和垂直于前一个投影的另外两个投影。在图像重建中，我们使用了有序子集期望最大化方法，并使用了上述四个子集。因此，即使只有12个投影数据，我们也可以获得与60个投影数据几乎相同质量的图像。

The purpose of our research is to develop a new data acquisition scheme for myocardial single photon emission CT with a semiconductor detector. A gamma camera with a semiconductor detector enables us to measure projection data with arbitrary angles because the gamma camera with this detector is smaller than ordinary gamma cameras with a scintillation detector. The proposed data acquisition scheme is to measure projection data with angles where the signal to noise ratio in projection data is relatively high. We assumed that we could measure the heart with 25 projections. In the x-y plane the positions of the detector were φ=-30, 0, 30, 60 and 90 deg. and they were 0=30, 60, 90, 120 and 150 deg. to the z-axis. Then we decided the minimum number of projections required for an acceptable quality of images, and selected optimum projection angles. The criterion for selecting the projection angle was an integral value of the power spectrum for each projection. We listed projection data in descending order according to this integral value and made four projection subsets. A subset was composed of three projections, i.e., the projection which had the largest integral value and the other two projections which were perpendicular to the former one. In the image reconstruction we used the ordered subset expectation maximization method and used the above four subsets. As a result, even with only twelve projection data we could obtain images of almost the same quality as those obtained with 60 projections.