Position mapping and a uniformity correction method for small-animal SPECT based on connected regional recognition

We describe a novel position mapping and the uniformity correction method to improve the performance of small animal single-photon emission computed tomography (SPECT) imaging. The SPECT system consists of a cerium doped lutetium−yttrium oxyorthosilicate (LYSO) scintillation crystal (22×22 pixel array, 2 mm×2 mm×3 mm pixel size), a position sensitive photomultiplier tube (H8500c, Hamamatsu Photonics Co., Ltd., Shizuoka Prefecture, Japan), and a parallel-hole collimator (Nuclear Fields Pty. Ltd., St. Marys, Australia). The position mapping method was based on a connected regional recognition algorithm. We present the algorithm and step-by-step details of image boundary detection, dynamic binarization, connected regional recognition, center-of-gravity computing, and look-up table establishment. The position mapping and uniformity correction tables were generated and applied to the SPECT projection data. The corrected projection images demonstrated that this correction method improved the uniformity of the raw projection image by ∼16%. The preliminary SPECT reconstruction results (using algebraic reconstruction technology) are also presented. A comparison between the reconstructed images before and after correction further confirms the performance of this correction method.