Noise Reduction Using Arithmetic Mean Filtering (A Comparison Study of Application to Different Noise Types)

Images are often corrupted by unwanted signals otherwise known as noise during acquisition and transmission alike, leading to loss of clarity of information in severe cases. Image restoration aimed at reduction in degradation and noise removal thus becomes imperative in digital image processing. This work focuses on the restoration of corrupted images in the presence of noise only. The arithmetic mean filter was applied to denoise an image sample corrupted by different noise types and its performance on the noise types was compared using the average percentage difference in the pixel values of the original and denoised image as well as the Peak-signal-to-noise ratio (PSNR). Simulation results show that the Arithmetic Mean Filter performs best on the image corrupted by Poisson Noise.