Gabor wavelet transform applied to feature extraction of ophthalmic images.



We investigate a new approach for measuring ocular

refractive errors (astigmatism, near-sightedness and shortsightedness)

from images of the human eye [1]. This

approach uses Support Vector Machines (SVMs) to

analyse data features extracted from images of the eye

acquired with a technique known as Hartmann-Shack [2]

and output measures for the eye´ s refractive errors. Prior to

analysis and feature extraction an image processing stage

takes place to identify and crop the region of interest

considering domain spatial and geometrical information.

Then a set of relevant image features are extracted using

wavelet and Garbor Transforms. These features are finally

fed into a SVM module for analysis that provides a

diagnosis of refractive errors present in the ocular globe

under investigation.