A Hybrid Self-constrained Genetic Algorithm (HSGA) for Digital Image Denoising Based on PSNR Improvement

The problem of denoising of images can be traced back to ancient times. Feature preservation remains an integral part of camera manufacturers around the world and drastic improvements have been achieved of late for the same. This work proposes a novel mathematical solution to the problem of image denoising. Images have been denoised using genetic algorithm evolutionary programming based on a self-constrained equational concept. A sample image is added with five different types of noise (Table 1) and they are denoised using existing filters (Table 2) and proposed algorithm (Table 3). The performance for different parametric functions has been compared using Peak Signal-to-Noise Ratio (PSNR) values in decibels. Consistent improvement is noted for five different noise models and compared with existing filters and the results are tabulated and graphically depicted (Figs. 4, 5, 6 and 7).