Document Type: Original Article
Research Center for Molecular and Cellular Imaging, Tehran University of Medical Sciences, Tehran, Iran
Department of Radiology, University of Massachusetts Medical School, Worcester, MA, USA
Research Center for Science and Technology in Medicine, Tehran University of Medical Sciences, Tehran, Iran
Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, USA
Department of Medical Physics and Biomedical Engineering, Tehran University of Medical Sciences, Tehran, Iran
Introduction: The position estimation in gamma detection system will have constant misplacements which can be corrected in the calibration procedure. In the pixelated crystal uniformly irradiation of detector will produce irregular shape due to position estimation errors. This image is called flood field image and is used to calibrate the position estimation. In this work we present a novel approach to automatically calibrate pixelated crystal array position estimation.
Methods: In the flood image of a pixelated crystal array the local peaks represent the estimation of position for the gamma photons that interacted in a single crystal pixel. First, the method detects 2-D peak locations automatically in the case of blurred pixel responses in the presence of noise and disturbance of the image. The algorithm consists of a filtering step for smoothing the image followed by two rounds of local peak detection. After localizing image peaks, the correction routine will map the image locations to the crystal pixels using the thin-plate spline interpolation method.
Results:The algorithm is tested for two flood images obtained from developed detector with different irregularity levels. By configuring constant parameters according to the detector configuration the method detected all crystal pixels in the image and map them correctly. The method further has been tested for10 identical blocks and the result showed automatic peak detection routine for all the blocks.
Conclusion: An automatic peak detection is presented to work instead of time consuming manual calibration routines. The method shows robust performance in the presence of image noise.