Indian Journal of Science and Technology
Year: 2016, Volume: 9, Issue: 32, Pages: 1-8
R. Rajkumar* and S. Saira Banu
Karpagam University, Coimbatore - 641021, Tamil Nadu, India; [email protected]
*Author for correspondence
Karpagam University, Coimbatore - 641021, Tamil Nadu, India;
Email: [email protected]
Objective: The Impulse or salt and pepper noise corrupt the image data and quality of the image. There are several strategies in the existing algorithms to eliminate the noise but also have drawbacks. We identify the drawbacks of the standard algorithms in consider with apply in the proposed algorithm to achieve the prefect noise removal. Enhance the Leading Diagonal Algorithm with different approaches to improve and identify the betterment of the proposed algorithm and prove with the parameter analysis of standard filters. Materials and Methods: Images are corrupted by the high density level of impulse noise. Retrieving the images by the proposed algorithm to make the images in better quality and preserving the edges. The Proposed algorithm based on the 5th pixel is called the pixel of processing and consider the diagonal matrix elements of the finding window to effectively to eliminate the high density noise. The Lena, The baboon and the lady-image are taken as the sample images to mix with the percentage of noises. The restoration of the sample images with the comparison of standard or recommended filtering algorithm. With the help of PSNR and MSE parameter to prove the improved algorithm is better than the available filtering algorithms. Findings: The noise density from 10% to 50% is added to the sample images in the increment of every 10% of noise density. Based on the parameter analysis (PSNR and MSE) with the standard filters to dedicate the new algorithm is for the betterment of more than the available filtering algorithm. The new algorithm is improving the quality and preserving the edges. Conclusion: To get the higher PSNR value and reducing value of MSE to prove the new approach is for the enhancement of the quality and retrieval of the images from the high density corrupted noise.
Keywords: Diagonal Sorting Method, Impulse Detection, Median Filter, Mean Square Error, Peak Signalto Noise Ratio
Subscribe now for latest articles and news.