Mammary gland segmentation using intensity correction for brightness attenuation in breast ultrasound image

This paper presents a method of segmenting mammary gland tissue for breast ultrasound images. In order to improve the quality of diagnoses for breast ultrasound screening, Computer Aided Detection (CADe) systems have been developed by a number of researchers. However, such systems suffer in terms of making many false positives for non-mammary gland tissue where no cancer exists. Accordingly, it is necessary to accurately determine the area of mammary gland tissue within ultrasound images for CADe systems. In this study, we propose a method of segmenting mammary gland tissue from ultrasound images for breasts. Within ultrasound images, brightness is attenuated as ultrasound waves move from the ultrasound probe to internal tissue. Accordingly, we correct intensity (brightness) using zero-phase component analysis (ZCA) whitening in order to reduce such brightness attenuation. Moreover, we apply a Conditional Random Field (CRF) which can render as smooth the region boundary of ultrasound images. The results of an experiment, conducted with five target patients to verify the effectiveness of the proposed methods, demonstrate that the method is capable of segmenting mammary gland tissue from breast ultrasound images.