Automatic Fascicle Length Estimation on Muscle Ultrasound Images With an Orientation-Sensitive Segmentation

The fascicle length obtained by ultrasound imaging is one of the crucial muscle architecture parameters for understanding the contraction mechanics and pathological conditions of muscles. However, the lack of a reliable automatic measurement method restricts the application of the fascicle length for the analysis of the muscle function, as frame-by-frame manual measurement is time-consuming. In this study, we propose an automatic measurement method to preclude the influence of nonfascicle components on the estimation of the fascicle length by using motion estimation of fascicle structures.

The method starts with image segmentation using the cohesiveness of fascicle orientation as a feature, obtaining the fascicle change by tracking manually marked points on the fascicular path with the Lucas-Kanade optical flow algorithm applied on the segmented image.

The performance of this method was evaluated on ultrasound images of the gastrocnemius obtained from seven healthy subjects (34.4 ± 5.0 years). Waveform similarity between the manual and dynamic measurements was assessed by calculating the overall similarity with the coefficient of multiple correlations (CMC). In vivo experiments demonstrated that fascicle tracking with the orientation-sensitive segmentation (CMC = 0.97 ± 0.01) was more consistent with the manual measurements than existing automatic methods (CMC = 0.87 ± 0.10).

Our method was robust to the interference of nonfascicle components, resulting in a more reliable measurement of the fascicle length.

The proposed method may facilitate further research and applications related to real-time architectural change of muscles.