Automatic extraction of subcutaneous fat from ultrasound images of cervical vertebrae

초록

Analyzing cervical vertebra is an important topic of caring patients with neck pains. While being noninvasive, cheap, and having real time convenience, ultrasound sonographic analysis has problems on subjectivity of inspectors. Thus, we need fully automatic assisting software for extracting major organs around cervical vertebrae. In this paper, we propose a method for automatic extraction of subcutaneous fat area with various image processing techniques. Carefully designed procedure including combining two types of binarization processes is the key point of the proposed method. Computationally heavy but sensitive adaptable fuzzy sigma binarization and simple thresholding binarization result are combined and the connectivity of the boundary lines is compensated by cubic spline interpolation. In experiment, the proposed method shows 97.6% success rate of extracting target area.

키워드