Hybrid filter for removal of speckle noise in digital images

Abstract

Speckle is a granular type of noise which degrades the quality of information (Synthetic Aperture Radar) SAR images and Medical images. So reduction of speckle noise act as a preprocessing step and it should be carefully designed to avoid spoiling of useful information. Adaptive filters such as Lee filter, Kuan filter, improved Kuan are some of the existing filters used to eliminate the multiplicative noise in SAR images and Medical images. A new method is proposed which uses the hybrid technique which involves the combination of Wiener filter along with the Adaptive Center Pixel Weighted Median Exponential Filter (ACPWMEF). The combination results in the proposed Hybrid Wiener Adaptive Center Weighted Median Filter (HWACWMF) are also determined. The proposed HWACWMF shows better results when compared to the ACPWMEF. The proposed filter is tested using the SAR images, Ultrasound (US) images and Magnetic Resonance Images (MRI). The simulation results are obtained using MATLAB R2013. The performance metrics are also measured and the quantity and quality of the images are improved using the proposed algorithm.