Image Noise Removal Using Principle of Suprathreshold Stochastic Resonance

In this paper, we have developed an algorithm for noise cleaning based on the principle of suprathreshold stochastic resonance. Stochastic resonance is the phenomenon in which the addition of right type and right amount of noise improves the detection of the signal in the system performance. In suprathreshold stochastic there are a number of array of detectors, all are subjected to the input signal and the same noise intensity distribution, then the noise-added signals are threshold, and a threshold noise added signal from each detectors are averaged to get the output image. We get the enhanced output when the noise-added input signal are threshold with respect to mean of the noise-added input signal. This algorithm was implemented on FreeMat open source software. PET scan of the Jaszack deluxe phantom image is performed and used as a input noisy image for the validation of the developed algorithm. The algorithm is successful in removing the noise from the image.