Development of Eye Movement Pattern Analysis System and Machine Learning Based Dizziness Diagnosis, Rehabilitation System

Abstract

The number of dizziness patients is increasing every year, and the age range is also varied. Therefore, there is a need for improved diagnosis and rehabilitation of dizziness with higher success rate. The diagnosis of dizziness is made based on the observation of the pattern of the nystagmus generated by the medical staff rotating the head of the patient. However, it is accomplished by the empirical and subjective judgment of the physician. Therefore, this study suggests an objective diagnostic algorithm based on artificial intelligence and a system for dizziness rehabilitation through HMD (Head Mounted Display) to overcome such misdiagnosis. The existing dizziness diagnostic equipment provides only the trajectory data of the horizontal and vertical movements of the nystagmus during the examination. However, this is only a supplementary information, and the medical staff reads the lesion directly based on the nystagmus image. However, the proposed system proposes three algorithms based on the optical flow technique that accurately track the three degrees of freedom motion including the rotational motion of the nystagmus. We also propose two diagnostic algorithms. One is a machine learning algorithm that analyzes the traced nystagmus data, maps the lesions on the 3D space and clusters them, and the other is a deep learning algorithm that learns nystagmus patterns to determine dizziness lesions. Finally, we will develop a rehabilitation system that will lead the dizziness patients to voluntarily manage dizziness treatment of dizziness and vestibular motion using HMD contents based system. This dizziness diagnosis rehabilitation system will be an indicator for the systematization and modeling of dizziness in the future. It is expected that it will be possible to use HMD for self-treatment as well as for small-sized and low-priced dizziness medical equipment, not only in the otolaryngology of small hospitals and private hospitals but also in individual households.