Automatic feature extraction and parameter estimation from unipennate muscles in B-mode ultrasound images

Abstract

Muscle ultrasound imaging is a convenient technique to visualize normal and pathological muscle tissue. Studying muscle fascicle pennation angle and fascicle length and their dynamic changes during contraction are important measures in skeletal muscle studies. In this thesis, we propose an automatic method that detects the pennation angle and the fascicle length of the vastus lateralis and the gastrocnemius muscle from B-mode ultrasound images. The method is based on the Radon transform for detection of aponeuroses and fascicle orientation. The performance of the method was tested on two different data sets, and produces mostly good results when the aponeuroses are detected correctly. However, there are some difficulties with the aponeurosis detection for the images where the aponeuroses were less prominent.