Robust thickness measure in ultrasound image sequences: A novel active contour tracking method for characterizing muscle fatigue

Muscle thickness (MT) measured from the ultrasound image sequences is an important determinant for understanding the contraction mechanics of muscle fatigue (MF). However, most of reports concentrate on manual or automatic methods using the prior knowledage of line-shaped aponeurosis and non-affine transformation. In this study, a novel active contour tracking method is proposed to realize a continuous and automatic MT measurement, of which procedure consists of image enhancement, keypoint tracking, aponeuroses segmentation and MT estimation. As a vital preprocessing step, regions of aponeurosis and non-aponeurosis involved with speckle noise can be respectively enhanced and suppressed. Subsequently, multiple pairs of keypoints are manually digitized at superficial and deep aponeuroses at the first frame, and contours of aponeuroses can be extracted using region-based contour model with global signed pressure and local correntropy based K- means. Finally, MT is geometrically obtained between intersection points from the lines of keypoints and boundaries of aponeuroses. Experiment results show that the proposed method here can provide a robust approach for estimating MT changes in MF analysis.