DWM (Directional weighted median) filter is very popular in filtering digital image and remove mixed noise. Fuzzy logic is implemented with median filters to improve its performance. In the previous work, fuzzy logic system is implemented with switching median filter and gives better performance than directional median filter as well as switching median filter. Experimenting directional median filter with same fuzzy logic system didn’t yield to better results therefore fuzzy logic parameters has been changes as per strong points of directional weighted median filter and a constant has been included in the filtering equation to improve the results. So in this proposed work, we have successfully implemented directional weighted median filter with fuzzy logic system which is proving better results than DWM and FSMF (Fuzzy Switching Median Filter). PSNR (Peak Signal to Noise Ratio)is used for qualitative analysis of results.

Abstract

Digital signal processing is most widely used to process the signal. In digital signal processing filters are used to remove some unwanted constituents from aspired signal. Windowing is a scheme of finite impulse response filters. Present paper proposes a new versatile window function. It has two variable parameters first one is window span N and another changeable parameter is r. when the value of variable parameter r increases width of major lobe of window also increases with better side lobe reduction and vice versa. Gaussian window and Kaiser window are the well-known variable windows. This paper shows that the proposed window has more desirable results in comparison of Gaussian and Kaiser window with low power loss and better side lobe reduction. To achieve minimum power loss peak side lobe level should have to minimum. Proposed window has low peak side lobe level (-17.681dB) in comparison of Gaussian (-11.836dB) and Kaiser window (-6.9704dB). Proposed work shows that the proposed window has finer spectral characteristic then Gaussian and Kaiser window. FIR filter formed by applying proposed window has narrow -3dB bandwidth (2π×0.320 rad/sample) corresponding to FIR filter formed by using Gaussian and Kaiser window. Ripple ratio of FIR filter plotted by applying proposed window (-144.321dB) is less corresponding to FIR filter delineated by using Gaussian and Kaiser which indicates that the proposed window will give better side lobe rejection and reduce the aliasing problem. In the biomedical field noise present in ECG signal can also reduce by using proposed window.

In this paper, we propose a new region-based Active Contour Model (ACM) that employs signed pressure force (SPF) as a level set function. Further, a flood fill algorithm is incorporated along with SPF function for robust object extraction. Signed pressure force (SPF) parameters, is able to control the direction of evolution of the region. The proposed system shares all advantages of the C-V and GAC models. The proposed ACM has an additional advantage i.e. of selective local or global segmentation. Flood Fill framework is employed for retrieving the object upon successful detection in the image. In addition, the computer simulation results show that the proposed system could address object detection within an image and its extraction with highest order of efficiency.

No abstract