Improving Breast Mass Visibility in Ultrasound Images using Spectral Analysis and Contrast Enhancement Algorithms

Rui Teixeira Ribeiro, Ruth English, and Julia Alison Noble


ultrasound imaging, spectral analysis, Medical Image Processing


In this paper, we present a method to estimate the degree of echogenicity for each image subregion across a breast ultrasound scan plane and use that information to generate a parametric image where the visibility of a mass is increased (particularly, increasing the confidence in identifying boundaries). For each image subregion, the frequency spectrum is computed and spectral peak values (SPV) are detected. Next, a partially-clipped quadratic transformation function controlled by statistical descriptors is applied to the spectral peak values and a color-coded parametric image is generated. A 20-case pilot study findings suggest that this type of parametric image adds useful information to the B-mode image for clinical assessment.

Important Links:

Go Back