Image analysis of the relationship between changes of cornea and postmortem interval

Abstract

Opacity of the cornea is one of the important indices for estimating time of death. Now the work is done by forensic medical experts. An unbiased estimation method is needed. This paper proposed a method for finding the relationship between changes of cornea and postmortem intervals by processing and analyzing images. Firstly, a histogram based image segmentation method is proposed to extract corneal regions from pictures of rabbit's eye. Secondly, texture and color features are used to describe the extracted corneal regions. Those features are carefully chosen to represent the changes of cornea in different postmortem intervals. A KNN classifier is used to reveal the association of image features and postmortem intervals. The experimental results show that cornea image features can be used to automatically estimate postmortem interval. © 2008 Springer Berlin Heidelberg.