Innovative Noise Reduction Strategies in Ultrasound Images Using Shearlet Transform and Bayesian Thresholding

Abstract

Uterine fibroids are prevalent benign tumors affecting women, often diagnosed through imaging modalities such as ultrasound. Ultrasound imaging is a widely used diagnostic modality for uterine fibroid due to its non-invasive nature. However, the images obtained often suffer from speckle noise, which can obscure fine details and complicate accurate diagnosis. Existing methods for removing speckle noise have limitations, including losing texture and edge information and not being able to handle low frequency noises. This paper presents a novel approach for speckle noise reduction by combining Shearlet Transform with Bayesian thresholding. The proposed method aims to achieve superior noise reduction while retaining important image features crucial for accurate diagnosis. Experimental results demonstrate the efficacy of the Shearlet Transform and Bayesian thresholding in significantly reducing speckle noise, enhancing image quality, and improving the interpretability of ultrasound images. Performance metrics like Mean Squared Error (MSE), Structural Similarity Index and Peak Signal to Noise Ratio (PSNR) helps to validate our proposed method. Reducing speckle noise in ultrasound images of uterine fibroids contributes to more accurate diagnosis and improves surgical treatment outcomes.