Wavelet based electroretinographic signal analysis for diagnosis

• To automate an ophthalmologic diagnostic system based on signal processing.
• Wavelet transform is used to extract the features.
• Entropy analysis is used and we have formulated an Electroretinographic Index to discriminate between controls and suspects.
• It will assist the physicians to cross check their diagnosis.

In this paper we have made a humble attempt to automate an ophthalmologic diagnostic system based on signal processing using wavelets. Electroretinographic signals indicate the activity of the retinal cells from different layers of the inner retina and therefore these signals are used to predict various dreadful diseases. In this work we have analyzed 95 subjects from four different classes viz. Controls, Congenital Stationery Night Blindness (CSNB), rod-cone dystrophy and Central Retinal Vein Occlusion (CRVO). The signal features extracted by wavelets are used for morphological and statistical analysis and for getting the subtle parameters like entropy. The results found comprises of difference in the values of wavelet coefficients, a-wave and b-wave amplitude in the case of normal and pathological signals. The colour intensity distribution of scalograms shows highlighting variations in the case of maximum response and oscillatory potentials of the ERG signals for specific type of diseases. Furthermore, we propose an Electroretinographic Index (ERI) from different entropy parameters which can be used to distinguish between the normal and abnormal classes. This new method based on ERG signal analysis can be reliable enough to build a solution for the constraints in the field of ophthalmology.