Eliminating Impulse Noise From Images By The Combination Of a Fuzzy Inference System and SWM -I Method

Impulse noise presents significant challenges in image processing, often resulting in degraded visual quality and the loss of crucial information. Common methods for detecting and removing noise include various approaches from the Switching Median filter (SWM) family. However, one method, SWM-I, is limited in its ability to identify all noisy pixels. This paper presents an effective approach for detecting and removing impulse noise by integrating a Fuzzy Inference System (FIS) with the SWM-I method. The proposed method is called FSWM. The SWM-I method is employed to effectively restore the corrupted pixels while preserving image details and edges. Subsequently, the FIS enhances the detection process by evaluating ۳x۳ and ۵x۵ windows that improve classification accuracy. Experimental results demonstrate that the proposed method significantly outperforms traditional noise reduction techniques, achieving superior performance metrics in terms of Signal to Noise Ratio (SNR), Root Mean Square Error (ERMS), and visual quality. This framework not only facilitates robust impulse noise mitigation but also provides a foundation for future research into advanced noise reduction methodologies.

Keywords:

Fuzzy inference system , Impulse noise , SWM method