Study of mixed filters schemes for denoising the medical ultrasound images

Abstract

The medical ultrasound images are usually tainted by a kind of noise called “speckle”. This noise is caused by the logical nature of the spreading occurrence. A simple hybrid scheme (combining Wiener and Mean filters) was designed in previous work [1] for denoising the medical ultrasound image. A performance comparison was made between the designed hybrid scheme and other filters (Wiener, Wavelet, Mean, Median and M3 filters) as stand alone filters. This paper presents our study which builds upon the above hybrid scheme by mixing filters (Mean, Median, M3, or the recommended Wavelet filter) with Wiener filter. These filters all have the window size 3∗3 and are completed by a designed simple adaptive filter. The performance comparison will be between the four mixed filter schemes above, to study their efficiency of denoising the medical ultrasound images. The version of the simple adaptive filter is dependent on the nonlinear gain function. The idea will be implemented by using the MATLAB program.