Medical Image Denoising Using BAT Optimization Algorithm

Abstract

Denoising is critical in medical imaging for the study of pictures, the diagnosis and treatment of illness. Image denoising approaches based on optimization are now effective, however the methods are constrained by the need for a large training set size (i.e., not successful enough for small data size). Medical picture denoising may be accomplished using the discrete wavelet transform (DWT) and a coefficient thresholding-based BAT method (CTB BAT). Denoising images by removing a residual from a noisy image yields denoised images, while most other image denoising methods start with latent clean images and work their way up to learning noise from the noisy images. Additionally, the wavelet transform is incorporated with CTB_ BAT to increase model learning accuracy and training time. Denoising strategies are compared to our model's performance in terms of peak signal-to-noise ratio and structural similarity in order to determine how well it performs compared to other medical picture denoising approaches. Our methodology outperforms other approaches in experiments, as shown by the findings.