Detection and Analysis of Gait Impairments in Lower Back Pain (LBP) Patients Using Image Processing Techniques

Every person has a unique gait, which is the pattern of individuals walking. Doctors can explore health conditions of an individual by studying stride, as many health issues reveal distinct gait manifestations. The posture of a person’s spine, neck, shoulder, and upper limbs has a big impact on gait. Lower back discomfort is a condition that significantly affects gait irregularity. Patients with persistent lower back pain (LBP) typically struggle to walk and move more slower than people who are healthy. Patients with LBP typically walk with more coordinated pelvic and thoracic spins than healthy individuals. In these cases, low thorax-pelvis relative phase seems to have an impact on arm swing. It is crucial to study the characteristics of gait in order to comprehend gait disorders. The currently used techniques use infrared cameras for motion detection, retroreflective tags, and micro electro mechanical sensors. The relevant literature is presented in this work, and the methodology adopted to achieve the goal of creating an image-processing technique-based non-invasive method to extract gait features is narrated. The procedure entails recording of the video with a mobile device and processing it for the extraction of the gait features. After the background is removed from the video frames, silhouettes are extracted. Tags are marked on the silhouette to understand and the characteristics of the human gait are measured. The dataset is created under medical supervision for normal and five forms of abnormal gaits namely spastic gait, scissors gait, steppage gait, waddling gait, and propulsive gait—six features relating to distance and angular measurements are retrieved. In this investigation, LBP was found to cause an aberrant waddling gait. The steps involved in the procedure are non-invasive, and the process of video recording needs no additional effort from the impaired person under test. The observed features in this study are found helpful for analysis of LBP. The developed procedure can be extended for analysis of arthritis and other disorders that cause gait impairment.