Robust Statistical Enhancement Techniques for High-Density Impulse Noise Reduction
Arinjay Bhowmick, Rudrajit Choudhuri, Amiya Halder · Advances in engineering research/Advances in Engineering Research · 2024
Image quality enhancement via impulse noise reduction is a critical phase in image preprocessing.Faults in the acquisition, storage, and transmission devices often corrupt the images by introducing noise that further hinders image analysis and processing tasks.This paper focuses on high-density salt and pepper noise removal from images using statistical image enhancement techniques.We present two enhancement algorithms targeted at noise removal achieved through pixel regeneration.The first approach uses two-stage filtration based on an adaptive substructure; the noise is primarily eliminated using the non-noisy neighbors in an adaptive window, followed by fine-tuning the pixel intensity to remove artifacts.The second approach uses a quasi-adaptive substructure where the neighbors in primary directions contribute to the decision-making process of pixel regeneration based on their information relevance.Performance evaluation based on the inferences made from different experiments on multiple images verifies the efficiency of the presented techniques.The observed reliability and robustness reflected in the results suggest the superiority of the algorithms over their existing peers.