An enhanced approach for vestibular disorder assessment

• Published in: 2018 IEEE 4th Middle East Conference on Biomedical Engineering (MECBME) pp. 243 - 246
• Main Authors: Ben Slama, Amine, Mouelhi, Aymen, Manoubi, Sondes, ben Salah, Mamia, Trabelsi, Hedi, Sayadi, Mounir, Fnaiech, Farhat
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-03-2018
• Subjects: Biomedical engineering; Classification algorithms; Conferences; Feature extraction; Image segmentation; Multi-layer neural network; Multilayer Neural network (MNN); Nystagmus; Principal component analysis; Principal component analysis (PCA); Vestibular disorder; Videonystagmography (VNG) system
• ISSN: 2165-4255
• DOI: 10.1109/MECBME.2018.8402441

Nystagmus is an abnormal eye movement symptom of vertigo related to a dysfunction in the peripheral vestibular system. The videonystagmography (VNG) is still inaccurate technique that suffers from many problems of insufficiency in distinguishing vestibular disorder diseases. In addition, doctors have poor information in the analysis of vertigo indicated by the involuntary eye movement. In this paper, a medical characteristics analysis method for angular displacement vectors of nystagmus is proposed. Firstly, the video images of nystagmus are captured with infrared video cameras and the motion trajectories of the eyeballs are obtained through pupil localization. Secondly, we compute the temporal and frequency variation of nystagmus in the two phases (fast and slow). Finally, principal component analysis (PCA) is used to select the pertinent component from all computed features as a preprocessing step for the classification stage of vestibular dysfunctions. Experimental results prove that the proposed method based on PCA method is very useful for an accurate way to evaluate the nystagmus condition and to remediate the problems of VNG technique such as erroneous information provided in the fast phase of the eye blinking movements. We also succeed to extract the most significant components from the temporal and frequency nystagmus features. Finally, we can say that the proposed segmentation preprocessing and classification method will provide high requirements for the diagnostic of the vestibular disorder datasets.