Anna Poghosyan

Applying Random Forests to the Classification of Forensic Evidence.

Started Mar. 25, 2022

Abstract or project description

Blood Spatters are one of the most common forms of forensic evidence in crime scene investigation. However, analyzing and classifying blood spatters takes a lot of time and energy, slowing down the investigation process. In particular, classifying the cause of impact stains as gunshot or blunt force could be sped up through the use of Artificial Intelligence. To tackle this problem, we used a publicly available dataset of 61 impact stain images and built a Random Forest Model that classifies the data as either gunshot or blunt force induced.