Processing Lenna Image Using Adaptive Weighted Median Filter

In this documentation, I processed the Lenna image using MATLAB with a custom implementation of adaptive weighted median filtering. The workflow began by reading the Lenna image from disk and converting it to grayscale format using MATLAB's imread() and rgb2gray() functions. I developed a custom function implementing the adaptive weighted median filter algorithm, which dynamically adjusts filter weights based on local pixel neighborhood characteristics. The algorithm calculates weight distributions by analyzing gray-level differences between the center pixel and its neighbors within a specified window size, allowing for more effective noise reduction in heterogeneous image regions. Key implementation aspects included: defining a sliding window mechanism using nested loops for pixel-wise processing, implementing dynamic weight calculation based on local gray-value variance, and utilizing MATLAB's median() function with weighted inputs. I optimized filter performance by adjusting parameters such as window size (typically 3×3 or 5×5) and gray-level difference thresholds through iterative testing. The processed image was saved to disk using imwrite() and displayed on screen with imshow(), demonstrating successful noise suppression while preserving edge details. This processing approach effectively enhanced Lenna image quality by reducing noise artifacts while maintaining important structural information.