Color impulse noise removal by modified alpha trimmed median mean filter for FVIN

• Published in: 2014 IEEE International Conference on Computational Intelligence and Computing Research pp. 1 - 8
• Main Authors: Yadav, Pranay, Singh, Parool
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2014
• Subjects: Color; Colored noise; Filtering algorithms; Fixed valued impulse Noise (FVIN); Image color analysis; Impulse Noise; Maximum likelihood detection; Mean Square Error (MSE); Noise measurement; Peak Signal to Noise Ratio (PSNR); Root Mean Square Error (RMSE); Unsymmetricness
• ISBN: 1479939749; 9781479939749
• DOI: 10.1109/ICCIC.2014.7238369

In this reserch article presents a novel method for the enhancement of color images, when images are corrupted by color impulse noise. According to planned algorithm the color noisy pixels are substituted by novel trimmed mean median value color images. Firstly, the color image (RGB) is sub-split up into three sections, i.e. Red (R), Green (G) and Blue (B) color pixel matrices, then all three matrices are checked for noisy pixels. In our proposed work divided in two parts. In first part is detection of noisy pixels and the second part is removal of noisy with details preservation like edges. When pixels values, are present in between 0's and 255's, it implies that they pixel are noise free pixels. Apply this scenario in whole color image pixels for the detection of color impulse noise. Second stage is the removal of noise. In this stage whole image is divided into a small 3×3 particular window and apply Unsymmetric condition in a small 3×3 window with the combination of mean median filter. Different color images are tested via proposed method. The experimental result shows better Peak Signal to Noise Ratio (PSNR) value, Mean Square Error (MSE), Root Mean Square Error (RMSE) and with better visual and human sensing. This method yields a better output for color impulse noise as compare to the other filters.