De-blurring Cardiac SPECT Images by Maximum Likelihood Approach

This paper presents a blind de-convolution algorithm for enhancing cardiac SPECT images by reducing the blur present in the image. The method is based on maximum likelihood estimate and in particular, the processing is done in a suitable color space. An iterative algorithm, without any prior information, is used to estimate the original image and the point spread function. Blur metric and peak signal to noise ratio are considered for performance evaluation of the algorithm. The effect of number of iterations on the quality of de-blurred image is also studied. Real medical images are used for appraising the algorithm