A Fast Balance Test Method of Screening for Vestibular Disorders Using Low Cost Camera and Personal Computer

• Published in: 2020 IEEE Eighth International Conference on Communications and Electronics (ICCE) pp. 382 - 387
• Main Authors: Dao, Viet Hung, Hoang, Thi Thu Hien
• Format: Conference Proceeding
• Language: English
• Published: IEEE 13.01.2021
• Subjects: balance test; Cameras; Microcomputers; Romberg test; Software; Software measurement; Streaming media; Stroke (medical condition); Testing; vestibular disorder
• ISBN: 9781728154695; 1728154693
• DOI: 10.1109/ICCE48956.2021.9352088

This paper presents a fast balance test method of screening for vestibular disorders. Vestibular disorders and their complications have become one of the most common health problems worldwide. Patients with vestibular disorders not only feel uncomfortable in daily living but also have higher risk of stroke. Because of the popularity of vestibular disorders, a good screening method can significantly save resources for hospitals and society. Currently, there are three vestibular disorder diagnostic methods including hearing tests, eye movement tests, and balance tests. The first two methods are sensitive and accurate. However, they require complicated and costly equipment. Among balance test methods, observing the body sway in several postures is a good choice when the testing time needs to be shortened; nonetheless, diagnostic results strongly depend on subjective evaluations of doctors. This paper contributes a solution by introducing a computer-aided method for quantitatively measure the degree of swaying. The proposed algorithm can process real time video captured by a camera, detect two colored stickers attached on the shirt, and calculate the inclination of the patient body. Many sub-processes have been utilized to enhance the measurement accuracy and method applicability. Experiments were performed with a mannequin and human bodies. Results show that angles measured by the proposed method and the reference data are almost equal. The plotted results show remarkably differences between healthy volunteers and patients. This means that the proposed system could provide a useful basis for screening and contribute an important step towards applying machine learning in vestibular disorders diagnosis.