Analysis of small tumours in ultrasound breast echography

Abstract

Breast ultrasound can image several different types of breast conditions, including both benign (non-cancerous) and malignant (cancerous) lesions. Ultrasound is frequently used to evaluate breast abnormalities and allows significant freedom in obtaining images of the breast from almost any orientation. Due to strong overlap of benignancy and malignancy in ultrasonic images and subjectivity of interpretation the digital processing of the images is a useful tool for an accurate and reliable diagnostic second opinion for physicians. This paper presents an analysis of small tumours of ultrasound breast using the methods: correction for non-uniform illumination, thresholding to create a binary image, counted method, measurement each tumour of ultrasound breast and build an array. It is concluded that a considerable amount of new, potentially useful clinical information can be revealed from ultrasound using image analysis approach proposed. These methods offer information about the characteristics of the small tumours and easily compute statistics for all the small tumours in the image.