COMPUTER NEURAL NETWORK FOR DIAGNOSTICS OF EYE DISEASES

glibdoadingAI and Robotics

Oct 20, 2013 (3 years and 9 months ago)

77 views


COMPUTER NEURAL NETWORK FOR DIAGNOSTICS OF EYE DISEASES


Victoria V. Dubich, PhD, Associate Professor

Krasnoyarsk State Technical University
, Russia

klimnatal60@yandex.ru


A computer neural network program is now developed for diagnostics of eye diseases

in ophthalmology. Nowadays ophthalmologists use a special device


reograph to
diagnose eye diseases, such as glaucoma, diabetic retinopathy and so on. A patient’s
eye is affected by weak electric current and the electrical resistance of eye tissues is
me
asured. The result of these measurements is a reogram


a periodic curve with
characteristic points. The elements of this curve contain information about eye
diseases. Reograms are used for developing a database for computer neural network
that can diagno
se eye diseases.


The first stage of work is input data preparation: analyzing the reograms and selecting
important parts, transforming the reograms into digital form, Fourier decomposition
of the curve, developing one database for the network training and

another one for the
network testing.


The second stage is training the computer neural network. Hundreds of reograms from
ophthalmology clinic with doctors’ diagnoses are used to train the computer neural
network.



The third stage of the work is testing

the computer neural network. The computer
neural network diagnoses are checked and compared with diagnoses of doctors.


The result of this work is going to be a trained and tested computer neural network
that will be able to diagnose diabetic retinopathy

with high accuracy.



Key words: ophthalmology, diabetic retinopathy, computer neural network,
Information technologies, diagnostics of eyes diseases.