A Differential Evolution Approach to PET Image De-noising
Vikas Gupta, Chi Chiu Chan, Pui Tze Sian · Conference proceedings · 2007
Positron emission tomography (PET) plays an increasingly potential role in modern clinical diagnosis. But, often, the quality of PET images is compromised due to the photon noise which reduces the signal to noise ratio (SNR) of these images. The degradation of images by noise may influence the diagnosis of critical diseases in clinical environments. The de-noising of PET images, by means of differential evolution based wavelet thresholding, is proposed here. This method keeps the resolution loss minimum and removes the photon noise from the images. The percentage of improvement in SNR was found to be more than 16% as opposed to 8% and 1.4% improvement obtained by means of median and Wiener filtering respectively.