Skip to main content

Using AI for landmark detection for facial reconstruction from images of skulls

Primary supervisor

Yasmeen George

Co-supervisors

  • Nick Dempsey

Problem Statement: The forensic identification of human remains is a critical legal process, culminating in the issuance of a death certificate by the appropriate authority. It is a multifaceted procedure that integrates scientific evidence—from antemortem records to advanced DNA analysis, forensic odontology, and anthropology—to match unidentified remains with missing persons. This holistic approach is foundational to resolving cases of unlawful killings, which bear significant implications for public health, legal resolution of civil affairs, and community well-being. Against this backdrop, AI with a forensic focus can potentially address the legal and humanitarian needs that arise during post-conflict periods and aiding the healing process of affected communities.

 

Student cohort

Double Semester

Aim/outline

In this project, we aim to build AI models that help identify missing persons from skulls. We will focus on training AI model for automatic landmark detection for facial reconstruction from images of skulls. 

Required knowledge

Python, ML libraries (such as sklearn), PyTorch, Image Processing libraries (such as OpenCV)