أكاديمي

Hewa Majeed Zangana (Duhok Polytechnic University, Iraq) and Firas Mahmood Mustafa (Duhok Polytechnic University, Iraq)

https://www.igi-global.com/chapter/a-novel-hybrid-wavelet-gan-image-denoising-system/377710

The hybrid approach offers two significant contributions; the first one by using wavelet trans-forms, the system achieves noise isolation and reduction across different frequency levels, ad-dressing both high- and low-frequency noise components effectively; and the second one the GAN framework introduces data-driven learning that enhances image details and restores subtle structures lost during the wavelet filtering process. Extensive experiments on various image datasets demonstrate that the proposed system outperforms conventional denoising techniques, such as traditional wavelet-based methods and pure GAN-based approaches, in terms of peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM). Furthermore, the hybrid system proves to be robust across diverse noise levels, making it a highly effective tool for real-world image denoising tasks.