The time of death is often to be estimated in forensic medicine to avoid wrong conclusions and to relieve innocently convicted persons. Due to cooling of the corpse during some hours after death, the body temperature can be used to estimate the time of death. Recent methods involve finite element simulations of the temperature distribution for the estimation. In this project we aim at improving the accuracy of time of death estimates by generating individual FE grids from CT scans, a more detailed modeling of the physical processes, and parameter identification methods also for uncertain parameters like heat transfer coefficients.

Foundation for the virtual corpse is a CT scan. The segmentation of the CT volume is performed using the ZIBAmira software. This results in distinct tissues of the human body being represented by 3D volume elements. The volume segmentation is used to generate a tetrahedral grid - the virtual corpse.

By individualizing the virtual corpse in size and posture with respect to a real corpse, it will serve as template for numerical computation of the cooling process described by the heat transfer equation.

Provided, we

  • know all the parameters in the heat transfer equation,
  • have an accurate approximation of the geometry of the corpse,
  • know the initial temperature distribution and the environmental temperature

then we can compute the evolution of cooling with high precision in every point in the corpse.

Having a measurement of the temperature in some point of the body, we could use the result of calculation to determine the time passed since death using the curve of temperature evolution in that point:

Unfortunately, the parameters in the heat transfer equation for a corpse are usually not exactly known. In order to identify the most influential parameters, we  compute the sensitivities, i.e., the partial derivatives of estimated time of death with respect to thermal tissue parameters (like heat capacity C, conductivity k, or heat transfer coefficient h). Thereby, we see that the tissues fat, muscle, and bone have the most significant influence on time of death, certainly due to their dominant volumes.