Poisson Noise Removal from Medical Images using Fractional Integral Mask

This paper uses a Fractional Integral Mask algorithm to remove Poisson noise from medical images. Riemann-Liouville definition of fractional calculus is used to create Fractional integral masks in eight directions. Two different methods of Mask Combining Technique, CT-1 and CT-2 are introduced for image de-noising. Performance of the algorithm is compared with that of Gaussian smoothing method of noise removal. Results depict that the algorithm with combining technique, CT-2 is better compared to CT-1. Experiments show that the mask size required directly depends on the fractional order. Mask size can be reduced for lower fractional orders thus ensuring the computation complexity reduction for lower orders. The operational range of fractional orders for CT-1 and CT-2 is also estimated. De-noising performance is measured based on visual perception and Peak Signal to Noise Ratio.