Microcomputer EOG-system For Ophthalmology
P.A. Semyonov · 2005
A completely automatic system for electrooculographic (EOG) investigations in ophthalmology is described. This system performs EOG-stimulation with investigation control and provides high accuracy results due to a unique real-time biosignal processing technique [ll, which reduces the operator 's involvement time up to 1/9-th of that for conventional EOG-technics. To optimize the processing technique, the EOG-biosignal's structure with the time and amplitude statistics of its components has been studied. The derived biosignal's additive model includes the EOG-informative and false components with the superimposed eye blinking, eye drift, head rotation and residual noise components as well as the high amplitude mains interference. The false and noise components can be expressed as the complex Poisson pulse series with the gaussian distributed amplitude, while the informative component is a quasiperiodical process with virtually unvaried amplitude. The biosignal real-time processing technique is based on the biosignal's model and performs the mains interference digital detection and removal, the adaptive quasi-optimal noise cancelling with optimal detection of the EOG-components, succeeded by the multifeature recognition of the EOG-informative component and its amplitude estimation. To minimize the demands of the processor's performance, the quasi-optimal noise cancelling and the EOG-components optimal detection procedures are combined using a time-moving un joined partially-linear approximation