PhD Defence Dimitar Rangelov | Enhancing Crime Scene Investigations | From Technology To Evidence

Enhancing Crime Scene Investigations | From Technology To Evidence

Dimitar Rangelov is a PhD student in the department Computer Sience. (Co)Promotors are prof.dr.ir. M. van Keulen from the faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS), University of Twente and dr. J.C. Knotter PhD from Saxion University of Applied Sciences and dr. R. Miltchev from Technical University of Sofia.

Crime scene investigation remains largely dependent on manual methods such as photography, sketching, and written descriptions. These are constrained by human perception, time pressure, and the impossibility of revisiting a scene once released. Meanwhile, 3D reconstruction, multi-modal sensing, and artificial intelligence have matured considerably, yet adoption is hindered by cost, insufficient training, institutional resistance, and a gap between public expectations and actual capabilities.

This dissertation investigates how these technologies can be responsibly and affordably integrated into crime scene investigation, so that they add evidential value without compromising accuracy, interpretability, or trust.

The technical core evaluates Neural Radiance Fields and Gaussian Splatting. Controlled experiments identify optimal camera settings and capture strategies, and a direct comparison shows Gaussian Splatting outperforming NeRF on every criterion, with lower noise, better detail preservation, and far shorter processing time. A CPU-based thumbnail generator achieves a 156-fold speed-up, giving investigators immediate previews without specialised hardware.

The research also examines AI as a decision-support layer, from spatial evidence contextualisation in firearms trafficking cases to systems generating plausible crime scene scenarios, and proposes an ethics-by-design framework with human-in-the-loop control, transparency, and auditability.

Three findings cut across the work. Accuracy in 3D reconstruction is conditional rather than absolute, so evidential reliability is inseparable from the transparency of the process that produced it. AI systems are interpretive mediators that encode assumptions, and uncritical acceptance risks automation bias. Trust requires institutional scaffolding, not technical performance alone. Treated as aids to reasoning rather than substitutes for it, these systems can counter tunnel vision and support transparent, evidence-based practice.