Machine Learning-Based Severity Classification of Spinal Cord Injury Patients Using Straight Leg Raising Test

In the world, approximately 250,000-500,000 new cases of spinal cord injury occur each year, and they require chronic treatment due to the intractable nature of the injuries. The mainstream method for evaluating trunk function in spinal cord injury patients is based on subjective evaluation by physical therapists. Because this evaluation is based on the experience of the physical therapist, the detailed evaluation method of trunk function is largely dependent on the skill of the physical therapist. Therefore, an objective and quantitative method of evaluating trunk function is necessary. This study focuses on the straight leg raising (SLR) test during rehabilitation. By measuring SLR using multimodal sensing and convolutional neural network (CNN), the severity of the patient's condition can be quantitatively assessed.