Adaptive and Optimized RGB Channel Filtering for Noise Reduction in Color Images

Abdelbary Algamel

doi:10.26629/jtr.2025.52

Authors

Abdelbary Mostafa Algamel Libyan Center for Engineering Research and Information Technology, Bani Walid, Libya , Pamukkale University , Department of Electrical and Electronics Engineering, Pamukkale University, Denizli, Turkey Author

DOI:

https://doi.org/10.26629/jtr.2025.52

Keywords:

Color Image Processing, Particle Swarm Optimization, Noise Reduction, RGB Filtering, Digital Image Enhancement, Adaptive Filtering.

Abstract

This paper introduces an adaptive filtering method for color image denoising that employs Particle Swarm Optimization (PSO) to tune filter parameters. Building upon a standard mean filter baseline, the approach uses PSO to independently adjust coefficients for each RGB channel according to the specific characteristics of both the image and the noise present. The optimization process targets dual objectives: minimizing Normalized Mean Squared Error (NMSE) and maximizing Signal to Noise Ratio (SNR), while maintaining the core properties of mean filtering. Experimental validation using controlled Gaussian noise demonstrates substantial performance gains: NMSE decreased from 4.81% to 1.11%, and SNR improved from 13.18 dB to 19.55 dB, representing approximately four-fold enhancement over conventional mean filtering. The method shows particular promise for applications demanding image specific optimization, though its effectiveness varies with image content and noise characteristics.