Generative Craniofacial Reconstruction as a Game-Changing Tool for Forensic Human Identification

Short description of the innovation study

Worldwide, sheer numbers of human remains are unidentified due to mass disasters, natural catastrophes, and violent conflicts. This is a dramatic challenge for forensic human identification techniques.

CranioFacial Reconstruction (CFR) can recreate the face of an individual solely from their skull, based on their strong morphological correlation. It is most useful when ante-mortem data is lacking, making other techniques unfeasible. To date, CFR relies on the scarce availability of expert forensic artists through a service that is costly, slow, and limited in access. 

The advent of generative Artificial Intelligence (AI) makes an automated CFR tool possible. To find the best approach to a functional CFR model, FACEGEN will develop specifically trained conditional generative adversarial networks, and fine-tuned foundation diffusion models. The resulting tool will provide an objective, reproducible methodology, enabling scalable business models and making forensic CFR widely accessible.

Organisations involved: 

ISV: Panacea Cooperative Research Soc. Coop.

HPC provider: Barcelona Supercomputing Center

Technology expert: National Institute for Research in Digital Science and Technology and University of Granada