Speckle Noise Removal from Three-Dimensional Ultrasound Images of Ovaries

Ultrasound imaging is often plagued by a plethora of noises, among which speckle noise stands out as an inherent characteristic, posing challenges in accurate diagnosis. This study aims to address this issue by proposing a composite filtering method tailored for de-noising Three-Dimensional (3D) ultrasound images of ovaries, which are particularly susceptible to speckle noise interference. The proposed filtering approach combines established filters with a novel filter termed the Modified Hybrid Mean-Median filter. Several prominent speckle noise filtering techniques including Wiener, Gaussian, Mean, Median, Hybrid Median, and the MHMM filter were applied to various slices of a 3D ultrasound image obtained through volume rendering. Their efficacy was then compared to ascertain the most effective filter for each slice. Subsequently, volume reconstruction was employed to generate de-noised 3D ultrasound images. Evaluation metrics such as Mean Square Error (MSE), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index (SSIM) were utilized to measure the performance of the different filters. The experimental setup utilized ultrasound images from USOVA3D Training Set 1, yielding promising results in terms of noise reduction and image quality enhancement.