Removal of fixed valued impulse noise by improved Trimmed Mean Median filter

Abstract

Impulse noise removal is considered one of the most burning topic in digital image processing (DIP). When an image is formed, factors like lighting (source, and intensity) and camera characteristics like the sensor response, lenses and also atmospheric condition affect the presence of the image. It hides the important fine points and information of images. In order to enhance the qualities of the image, the removal of noises becomes imperative and that should not be at the cost of any loss of image information like edges. Removal of noise is one of the most important pre-processing tasks of different of image analysis works and tasks like image enhancement, steganography, segmentation and other enhancement related process. In this research article, we have proposed a new method for the removal and restoration of gray images is introduced, when images are corrupted by impulse noise. This method proposed a novel combination of Mean. Median and trimmed value concept for elimination of fixed valued impulse noise. Our methodology ensures a better performance for different level low, medium and high density of fixed value impulse noise as compare to the other famous filters like Standard Median Filter (SMF), Decision Based Median Filter (DBMF) and Modified Decision Based Median Filter (MDBMF) etc. The main objective of the proposed method is to improve not only a peak signal to noise ratio (PSNR) but also improve the visual perception and reduction in blurring of the resultant image. In the proposed method when previous pixels values, 0's and 255's are present in the particular window and all the pixel values are 0's and 255's then the remaining corrupted pixels are substituted by mean and median value. Proposed methodology was tested on gray-scale images like Mandrill and Lena. The experimental result shows improved value of peak signal to noise ratio (PSNR) and mean square error (MSE) values with better visual and human perception.