Blood Vessel Segmentation

Motivation

Glaucoma is one of the various eye-related conditions that require early treatment. It is a manifestation of a group of Conditions originated from different causes resulting in an increased pressure inside the eye. Often, if defects were to be detected, they would be too late as significant damage to the nerve fibres may have occurred, causing a certain level of visual field loss. Therefore, as there is no cure for glaucoma, early detection of glaucoma is important because it can minimise damage and vision loss and allow for prompt and adequate treatments in avoiding blindness.

ONH is the main marker to detect Glaucoma, but the extraction of ONH from fundus images is not accurate due to the presence of Blood vessels and other fundus pathogens. So in this study Blood vessels segmentation is the first step. I have applied Active contours(snakes) for this purpose.

For segmenting the blood vessels we have used a deform able model based approach known as Active contours. This model is more generic and robust as it's performance is unaffected by the blood vessels inheriting properties such as vessel crossing and branching and Central Vessel Reflex problem. It also provides the details of the vessel widths and is having greater scope to develop.

Active contour model represents an object boundary or some other salient image features as a parametric curve (which is referred as snake). An Energy function E is associated with the snake is defined as the sum of

1. Internal energy
2. External energy
3. Image forces

The problem of finding object boundary is cast as an energy minimization problem. Internal energies give the model tension and stiffness whereas External energies come from high-level sources such as human operators or automatic initialisation procedures. Image energy is used to drive the model towards salient features such as light and dark regions, edges, and terminations.

The configuration of the snake is driven by the negative energy gradients which are referred as the forces. Using of Gradient Vector Flow method to determine the energy forces reduces the probability of falling into local minima. Active contour model for the segmentation of blood vessels doesn’t include internal energy as the vessel shape is highly torturous. So far three external energies affecting the snake were considered which are

1. Edge distance energy: Edge distance energy helps the snake advance of nodes close to vessel boundaries but it also stops them when they reach a minimum, that is, when they reach an edge point
2. Crease distance energy: It corresponds to the creases distance energy that drives the snake along the arteriovenous structure and blocks it if a maximum distance threshold is reached. and
3. Inflation force: It is a force without energy term and is the strongest expansion force which expands the snake subject to certain constraints. It is calculated once the snake has been initialized.

We have implemented the snakes on few samples from the DRIVE database available. We were able to extract the creases and initialize the snakes onto the blood vessels for segmenting them and were successful.
Segmented vessels for a sample image in the drive database is shown below and results for other images in the drive database can be found in the Report