Abstract

Nuclear medicine is playing a major role in medical diagnosis. But, due to limitations in both injected radionuclide to the patient’s body and low count rates of the detector, output images are of very low contrast. Several methods have been proposed to improve the contrast of medical images. In this study, a new method is presented for SPECT images. The proposed method is based on the combination of gradient approximation (GA) and histogram equalization (HE) algorithms to improve the image contrast. Poisson editing concept is deployed to allow the images to be edited and processed in the gradient domain before the reconstruction phase. GA is initially applied on the images to overcome the limitations of HE method. Using the GA concept, image gradients are manipulated first and then the images are reconstructed. These reconstructed images are fed as input for the HE block. Finally, results are presented both qualitatively and quantitatively.