Microcalcification enhancement in ultrasound images from a concave Automatic Breast Ultrasound Scanner

In this paper, we have proposed a tensor-based filtering method to enhance the SNR of hyperechoic spots in human breast ultrasound images while suppressing the speckle noise. The breast images were acquired by an Automatic Breast Ultrasound Scanner (ABUS), which offers high spatial resolution and the fast full breast exam capability. The enhanced images were then gone through a pattern recognition kernel to distinguish the small hyperechoic spots against large tissue boundaries. The isolated hyperechoic spots were finally merged back to the original image. We believe the proposed method can facilitate the detection and analyzing of microcalcifications (MC) in the noisy ultrasound images. This could offer a non-invasive alternative solution for the mammography in breast cancer diagnosis.