Automatic classification of muscle condition based on ultrasound image morphological differences

Abstract

Myofascial Pain Syndrome is a form of chronic muscle pain centered on sensitive points in muscles called trigger points. These points are painful when pressure is applied on them and can produce referred pain, referred tenderness, motor dysfunction and autonomic phenomena. Currently, the location of trigger point is mostly determined through physical examination by clinicians, which is considered unreliable due to the dependency on the clinician’s discretion. This study had developed a system that quantifies the location of trigger point using ultrasound images to detect the presence of trigger point. Normal muscle and muscle with trigger point shown morphological difference in ultrasound images,in which, is accentuated through image processing and pattern recognition. Statistical properties of the final signal output were analyzed to determine the most optimum value used for classification. Two parameters were calculated which are the mean and the standard deviation. Upon observation, the value of standard deviation can be used in setting the threshold value for the classifier to differentiate between normal muscle and muscle with trigger point. Based on the results, classifier can be set between 9 to 12 for DUS 100 and 13 to 19 for Aplio MX in order to successfully classify the images. System performance testing shows that this system has high accuracy when detection was performed with the current collection of sample images.