Scapular Dyskinesia Detection using Machine Learning

Scapular Dyskinesia is a disorder in which scapula performs an unusual movement during movements of the upper limb. The disease is usually associated with pain in shoulders, decrease in strength and limitation of functions in daily routine. Early identification of Scapular Dyskinesia is crucial to avoid future complications and successful rehabilitation. However, most methods for diagnosing rely on clinical observation by an expert. In addition, several studies have shown that some of the current algorithms that allow automatic diagnosis need big datasets and lots of computational power. In this paper, we will demonstrate our solution for automatic scapular dyskinesia detection based on machine learning models and video data analysis. Our method is based on video records that undergo analysis using OpenCV library to calculate necessary scapular features. The data is later used to train the machine learning classifiers that decide if the scapular dyskinesia occurred. Our method will aim to lower the computational load required for accurate detection. It will also provide a solution to some practical problems, such as lack of a large training set, differences in postures, lighting and others. The system has been developed to act as an aid to doctors and physiotherapists in early diagnosis and evaluation, with the aim of improving the quality of patient care in the long run.