Machine Learning Approach for Comparative Analysis of De-Noising Techniques in Ultrasound Images of Ovarian Tumors

Abstract

Ovarian abnormalities such ovarian cysts, tumors, and polycystic ovaries are one of the serious disorders affecting women's health. In ultrasound imaging of ovarian abnormalities, noise during capturing of the image and its transmission process frequently corrupts the image. In order to make the best judgments possible at the appropriate moment, ovarian cysts in females must be accurately detected. In computer aided diagnosis of ovarian tumors, preprocessing is a very important step. In preprocessing, de-noising of medical images is a particularly a difficult task since it must be done while maintaining image features that are essential for diagnosis. In this research work we are using various denoising filters on ultrasound images of ovarian tumors. For different noise denoising techniques, performance measures like MSE, PSNR, SSIM, and UQI etc. are calculated. According to experimental findings, Block matching 3-D filter outperforms all other methods. Radiologists can better diagnose the condition with the use of this computer-assisted system.