MOTF: Multi-objective Optimal Trilateral Filtering based partial moving frame algorithm for image denoising

• Published in: Multimedia tools and applications 2020-10, Vol.79 (37-38), p.28411-28430
• Main Authors: Rejeesh, M R, Thejaswini, P
• Format: Article
• Language: English
• Subjects: Algorithms; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Image filters; Image reconstruction; Matching; Multimedia Information Systems; Multiple objective analysis; Noise; Noise reduction; Optimization; Random noise; Signal to noise ratio; Special Purpose and Application-Based Systems
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-020-09234-5

In this paper, a novel denoising approach based on optimal trilateral filtering using Grey Wolf Optimization (GWO) is proposed. At first, a database of noisy images are generated by adding Gaussian noise, Salt & Pepper noise and Random noise to the captured image. The filtering of noisy images are performed by Block-matching and 3D filtering (BM3D) algorithm over the components of image obtained through the moving frame approach. Then, using optimal trilateral filtering, the denoised images are reconstructed. Therefore, by using a two-level filtering approach such as Moving frame-based Block-matching and 3D filtering (BM3D) and Optimal trilateral filtering the noisy images are decomposed. The proposed optimal trilateral filter employs Grey Wolf Optimization algorithm for selecting the parameters optimally to improve the efficiency of filtering method which also reduces the time required for manual computation. The performance of the proposed image denoising algorithm is analyzed using multiple datasets and the analysis of results were done in contrast with existing conventional approaches. The results validated that the optimal trilateral filtering approach outperforms other conventional methods in terms of Mean-Square Error (MSE) and the Peak Signal-to-Noise Ratio (PSNR).