Efficient Implementation of An Improved Median Filter on TMS320C6416 Digital Signal Processor

This paper proposes a new software design and hardware implementation to enhance grayscale images that are highly corrupted by impulse noise. However, the main objective is to remove the salt and pepper impulse noise using an improved median filter that performs spatial information processing to determine which pixels in an image have been affected by impulse noise then replace it only by the median value of the 2-D selected area which has low variance value. The proposed algorithm was optimized and implemented on a fixed-point TMS320C6416 Digital Signal Processor of Texas Instruments. The improved median filter (IMF) is successfully tested with multiple grayscale images and gives better Peak Signal-to-Noise Ratio (PSNR) results than the Standard Median Filter (SMF), Recursive Median Filter (RMF), adaptive Median Filter (AMF).