Background: Mass fatality incidents involving highly fragmented human remains (FHRs) represent major challenges for Disaster Victim Identification (DVI) units. Conventional recovery and identification protocols are frequently destructive, time-consuming, with a significant financial cost. Through digitizing the FHRs and the spatial data of the disaster scene, virtual anthropology and forensic archaeology enable the development of computer-aid workflows for the virtual re-association(VRA) of the FHRs, and for reconstructing their dispersion on scene. This integrative approach would allow for a better data preservation and traceability, an efficient identification of the FHRs and the analysis of the fragmentation mecanisms. This approach however necessits the standardization and scientific validation of the proposed protocols.
Objectives: This thesis aimed to (1) develop and assess a standardized virtual re-association protocol tailored for DVI constraints, (2) integrate VRA with georeferenced recovery methods to preserve and analyze spatial context of the scene, and (3) propose the Virtual Bonescape framework as an integrative approach, to reconstruct both victim identity and event dynamics.
Methods: Experimental studies were performed using porcine models to simulate high-fragmentation disaster scenarios (controlled explosion and random scattering). Computed Tomography (CT)-scan imaging was acquired to create digital models for the VRA of the fragments. Two spatial documentation strategies - archaeological grid and drone-based photogrammetry - were assessed and compared for their accuracy, efficiency, and operational feasibility.
Results: VRA was confirmed as a robust and feasible protocol, that offers significant advantages over traditional physical re-association (PRA). The preservation of FHRs spatial data using the archaeological grid proved more suitable for degraded conditions, while drone-based mapping proved faster and more scalable. For both approaches, the integration of georeferenced data with VRA enabled the conceptualization of the Virtual Bonescape, a novel framework that reconstructs the dispersion of fragments on scene and supports the retrospective interpretation of events.
Conclusion: This thesis highlights the potential of virtual anthropology and forensic archaeology to broaden the scope of forensic analysis beyond the sole identification of FHRs in mass disaster contexts. From a scientific standpoint, it confirms the relevance of the VRA protocol and demonstrates the feasibility of its integration with georeferenced spatial data. From an operational point of view, it underscores both the strengths and limitations of its implementation in forensic contexts. From an ethical perspective, it highlights the significance of identifying FHRs, using techniques that are both technically and financially accessible, as well as the added value of post-event scene analysis for a deeper understanding of the disaster circumstances. This research opens up multiple avenues for further research, including the use of artificial intelligence for automation of the VRA protocol, the development of predictive dispersion models, and the application of the Virtual Bonescape framework to contexts such as terrorist attacks and post-conflict reconstruction processes.
