Blood Vessel Detection from Sclera Images

EasyChair Preprint 1777

Abstract

After segmentation of the sclera area from the eye image, the next step for recognition of an eye is extraction of blood vessels within sclera, which do not change over time and is unique to every person. The method does not search for red patches on the sclera, but looks for line patterns on the sclera which are blood vessels, which is an essential requirement, as red patches for two eyes may appear similar under certain conditions. Contrast limited adaptive histogram equalization has been used to enhance the image. Images of detected vessels of two different images of same eye look very similar, even under low illumination. A trivial comparison logic is able to compare all images with minimal false accepts. Since the method is independent of colour of the vessels, it can be easily adapted to most real life situations.

Keyphrases: Eye Recognition as a Biometric, Image Morphology, Sclera Blood Vessel Detection

@booklet{EasyChair:1777,
  author    = {Sumanta Das and Ishita De Ghosh and Abir Chattopadhyay},
  title     = {Blood Vessel Detection from Sclera Images},
  howpublished = {EasyChair Preprint 1777},
  year      = {EasyChair, 2019}}