< Back to previous page

Project

Forentrics: Automated Sensing and Analysis of Crime Scene Evidence

Every year, hundreds of people are convicted of crimes they did not commit. Major contributors to such miscarriages of justice are bias, prejudice, junk science, and contaminated evidence. The National Academy of Science (NAS) published a milestone report in 2009, in which it explicitly called for the development of automatic and objective methods. In light of this report we developed HemoVision, a desktop application that uses machine learning and computer vision paradigms to automate the process of bloodstain impact pattern analysis completely. HemoVision has strong commercial potential but is still a few steps away from becoming a complete product. With this project, our aim is to develop and integrate strongly needed functional upgrades, validate and disseminate our results, and develop a complete business model. In doing so, we aim to contribute to the changing field of forensics and reduce the chances of judicial failing.
Date:1 Oct 2020 →  30 Sep 2022
Keywords:automated analysis, objective results, judicial failing, forensic science, crime scene investigation
Disciplines:Forensic medicine not elsewhere classified, Machine learning and decision making