2D ultrasound image segmentation using graph cuts and local image features

Abstract

Ultrasound imaging is a popular imaging modality due to a number of favorable properties of this modality. However, the poor quality of ultrasound images makes them a bad choice for segmentation algorithms. In this paper, we present a semi-automatic algorithm for organ segmentation in ultrasound images, by posing it as an energy minimization problem via appropriate definition of energy terms. We use graph-cuts as our optimization algorithm and employ a fuzzy inference system (FIS) to further refine the optimization process. This refinement is achieved by using the FIS to incorporate domain knowledge in order to provide additional constraints. We show that by integrating domain knowledge via FIS, the accuracy is improved significantly so that further manual refinement of object boundary is often unnecessary. Our algorithm was applied to detect prostate and carotid artery boundaries in clinical ultrasound images and shows the success of the proposed approach.

Simultaneous Lesion Segmentation and Bias Correction in Breast Ultrasound Images

“…Several segmentation methods have been applied to breast US images, from the most traditional histogram thresholding methods [5,11] to novel approaches based on graph-cuts [7,12]. Although active-contours methodologies are widely used [6,8] to determine the outline of the object of interest, they fail when dealing with blurred boundary lesions.…”

Evaluating Lesion Segmentation on Breast Sonography as Related to Lesion Type

“…Several segmentation methods have been applied to breast sonograms. The most used techniques are histogram thresholding, 6 – 8 active contours, 9 – 15 neural networks, 16 – 18 graph cuts, 19 , 20 and Markov random fields (MRFs) 21 , 22 . Histogram thresholding methods remove speckle noise or enhance lesion structures, obtaining satisfactory results in segmenting lesions with homogeneous content such as cysts, but it is often unreliable for segmenting cancerous lesions, which usually have a high variance of pixel intensities within the lesions.…”