Data Analysis Of Nystagmus

This paper aims to detect the vestibular function through the analysis of the dynamic eye information. Nystagmus data analysis is one of the objective bases for the vestibular function. The main method of this paper is to process the eye movement video images and then obtain displacement information of the pupil which can help to draw nystagmus wave curve and analyze the nystagmus wave curve.The content of this research includes the following five aspects:1)Collecting the displacement information through tracking the pupil by video. The Canny algorithms can be used to detect the edge of the image in this study, this algorithm can be used to detect the image boundary by two thresholds, the calculation is highly efficient and the better robustness.2) Person's pupil is similar to circular, so the pupil in the photo is detected by using detection of circle of the algorithm of Hough transform, the coordinates of pupil center and center point shift are calculated by adding the edge pixels to fitting circle.3) This paper uses the video tracking method which based on the optical flow to follow the pupil in the eye movement video, which use the corner detection method based on edge.4) This article stores the displacement information to the database, it use ADO technology to connect the access database.5) Three-tier-feed-forward neural network BP algorithm takes effect in curve fitting with the nystagmus data. In order to prove the data which is processed follow Gaussian distributions, it use histogram.In addition, it also recognizes nystagmus rapid and slow phase.This paper analyzes and processes the two groups of videos that smoothly tracks the eye movements and eye vibration with temperature,effectively achieving the eye vibration data,improving the method above right,eliminating the abnormal data,making it in the-80 to 80 effective interval and raising the accuracy of curve fitting.The eye movement information is obtained by dealing with the video images of eye movement, which can avoid electromagnetic interference in the electronystagmography. It is shown through experiment that the movement locus of pupil can be recorded effectively using the tracking algorithm, nystagmus information can be presneted directly by the wave curve.