Retinal vessel segmentation using the 2-D Gabor wavelet and supervised classification

Abstract

We present a method for automated segmentation of the vasculature in retinal images. The method produces segmentations by classifying each image pixel as vessel or nonvessel, based on the pixel's feature vector. Feature vectors are composed of the pixel's intensity and two-dimensional Gabor wavelet transform responses taken at multiple scales. The Gabor wavelet is capable of tuning to specific frequencies, thus allowing noise filtering and vessel enhancement in a single step. We use a Bayesian classifier with class-conditional probability density functions (likelihoods) described as Gaussian mixtures, yielding a fast classification, while being able to model complex decision surfaces. The probability distributions are estimated based on a training set of labeled pixels obtained from manual segmentations. The method's performance is evaluated on publicly available DRIVE (Staal et al., 2004) and STARE (Hoover et al., 2000) databases of manually labeled images. On the DRIVE database, it achieves an area under the receiver operating characteristic curve of 0.9614, being slightly superior than that presented by state-of-the-art approaches. We are making our implementation available as open source MATLAB scripts for researchers interested in implementation details, evaluation, or development of methods.

Retinal Blood Vessel Segmentation Using Gabor Wavelet and Line Operator

“…5, October 2012 appear brighter than the background is used as input image to our vessel detection system. An iterative algorithm as proposed in [1] is used to remove the strong contrast between the retinal fundus and the region outside the aperture. This improves the undesired response of both the wavelet transform and the line operators at border of retinal disk.…”

“…A family of wavelet can be defined by translations, rotations and dilations of the analyzing wavelet. The continuous wavelet transform is defined in terms of the scalar product of f with the transformed wavelet [1], [6]:…”

“…In [1] a Bayesian classifier in [10], [11] and [12] a K-Nearest Neighbor and in [2] Support Vector Machine have been used as classifiers for pixel segmentation. In classification process each pixel of retinal image takes one of two labels: vessel or non-vessel.…”

“…In [1], a method based on multi-scale Gabor analysis of retinal image was proposed. A feature vector consist of five features was used for vessel segmentation.…”

“…We use a feature vector consist of seven components: the grey level of green channel, response of Gabor filter bank [1] (four scales) and the response of two line operators. We use Gaussian mixture model to estimate class conditional density of vessels and non-vessels.…”