PU scholars develop AI-based facial reconstruction models

Panjab University scholars have developed AI-based facial reconstruction models that can accurately recreate a person’s face (with up to 95 per cent accuracy) using only the dimensions of the teeth and jaws — which could help in identifying victims in airplane crashes, bomb blasts, and other disasters where the face is damaged but jaws remain intact.

The innovation is expected to significantly strengthen forensic investigation and disaster victim identification (DVI) processes.

The university has received a copyright for the novel computer program from the Copyright Office, Government of India. The AI/ML models were devised by a team of forensic scientists under the leadership of Professor Kewal Krishan, Department of Anthropology, which included Professor Vishal Sharma (Institute of Forensic Science and Criminology – IFSC) and Dr Arun K Garg (Dr Harvansh Singh Judge Institute of Dental Sciences & Hospital).

As many as six PhD research scholars — Damini Siwan, Nandini Chitara, Ankita Guleria, Rakesh Meena, Akansha Rana, and Ayushi Srivastava — contributed to the project.

The models are capable of assisting forensic anthropologists in cases where human bodies are dismembered, decomposed, or commingled — conditions commonly found in natural and man-made disasters. With the AI tool, even when the face is damaged beyond recognition, identification becomes scientifically possible using preserved dental and jaw structures.

The research is part of the PhD work of Damini Siwan and Dr Nandini Chitara. Damini, a forensic scientist from IFSC, focuses on applying AI to forensic casework, research, and ethics. Dr. Nandini Chitara, recently awarded her PhD, is currently serving as Senior Scientific Assistant (Biology) at the Regional Forensic Science Laboratory, Haryana, Panchkula.