Adaptive Weighted Mean-Median Filtering for Robust Salt-and-Pepper Noise Removal Technique

The primary challenge with image processing applications in automated surveillance, medical, and remote sensing is image denoising. Salt-and-pepper noise (SAPN) drastically reduces image quality by randomly changing pixel values with high intensities. At higher noise densities, the fundamental challenge for conventional filtering algorithms is to balance noise suppression and detail retention. In digital image processing applications accuracy is very important. However, during capturing and transmission, the images are exposed to various noise frequently. In this research article, an Adaptive Weighted Mean-Median Filter (AWMMF) is proposed for robust Salt-and-Pepper Noise Removal Technique. In the proposed work the filtering window size is dynamically adjusted according to the local noise density. AWMMF integrates a weighted combination of mean and median values to enhance restoration quality while preserving image details. The efficacy of the proposed algorithm is evaluated on standard benchmark Lena image and compared with existing denoising techniques like Adaptive Fuzzy Median Filter, Fast and Efficient Median Filter, Nonlinear Hybrid Filter, Improved Adaptive Type-2 Fuzzy Filter, Regeneration Filter, Deep Convolutional Neural Network and Adaptive Switching Modified Decision-Based Unsymmetric Trimmed Median Filter. For the performance analysis, the parameters considered are the Peak Signal-to-Noise Ratio, Mean Squared Error, Structural Similarity Index and Image Enhancement Factor. AWMMF provides a robust and computationally efficient solution for SAPN removal, making it suitable for real-world image processing applications.