Implementation of edge preserved denoising technique for impulse noise removal in images
J. Praisy Faith, S. Priyadarshini, Adarsh Vijayan Pillai · 2017 IEEE International Conference on Power, Control, Signals and Instrumentation Engineering (ICPCSI) · 2017
In many industrial and medical applications, during the image acquisition, transmission and signal conversion, the image gets corrupted with impulse noise. For a clear and pleasing visual appearance of the image, a “Simple Edge Preserved Denoising Technique” (SEPD) is proposed to remove the impulse noise in the image. This technique achieves excellent tradeoff between noise suppression and detail preservation. In this technique, only the noisy pixels are manipulated and the other noiseless pixels are retained as it is. By applying the Edge Detection algorithm the edges and the fine details in the images are preserved. For each window, the edges are mapped in all the possible directions. The differences between the edges are calculated and the minimum value is found. The mean of this minimum valued edge is computed and is compared with the pixel to be manipulated. If this value is within the threshold limit the pixel is kept unchanged. A reliable threshold limit has been chosen. Due to the less computational complexity, the area is minimized and the speed is increased. Thus it provides an efficient VLSI implementation. This algorithm is applied to images with different noise ratio and the results are obtained.