Improved Decision Based Filtering algorithm for impulse noise removal in Digital Images

Abstract

Random valued impulse noise is removed in digital images, and an Improved Decision Based (IDB) filtering technique is proposed. The decision-based filtering algorithm’s function is to detect noisy pixels by making judgments in three different modules. Once the noisy pixel has been found, the noise is removed using a Hybrid Wiener Adaptive Centre Weighted Median filter. Comparing the existing decision tree-based approach, the suggested filter removes substantially more noise. The suggested modified decision tree method produces better outcomes for high-density noise levels of up to 90%, according to both visual and quantitative data.