Abstract
The identification of accident victims who arrive at hospitals unconscious, severely injured, or without identity
documents can be difficult when information is held in separate institutional records and identification depends on
manual checks. This paper presents VeilFind, a proposed and implemented web-based prototype intended to support
biometric identification and case coordination among citizens, hospitals, police, and system administrators. The
platform is designed around role-specific dashboards and workflows for citizen enrolment, document submission,
hospital-side victim record capture, police accident reporting, and administrative management. Its technology stack,
as documented in the project report, includes React, TypeScript, Tailwind CSS, Vite, Supabase, and GitHub. The
design also describes an AI matching capability based on facial and document embeddings, with match confirmation
intended to inform subsequent improvement of the matching process. The report documents functional test cases
covering registration and role assignment, biometric enrolment, hospital matching, accident logging, administrative
approval, role-based access, data privacy, and responsive presentation; all are recorded as passed. However, detailed
test logs, independent accuracy measurements, performance benchmarks, formal user-study data, and live integration
with national registration, hospital, or police production databases are not provided. The work is therefore best
understood as a software prototype demonstrating coordinated workflows and an AI-assisted identification concept,
rather than a validated national biometric service.
Keywords: accident victim identification, biometrics, facial recognition, artificial intelligence, Supabase, role-based
access control, emergency response, Malawi.
1. INTRODUCTION
The timely identification of people involved in road traffic accidents and other emergencies is important for
emergency care, family notification, police investigations, and related legal or administrative processes. When a
Citation
Atupele Ajay Bakili, MR. MTENDE MKANDAWIRE (2026). DESIGN AND IMPLEMENTATION OF VEILFIND: AN AI- ASSISTED DIGITAL SYSTEM FOR BIOMETRIC IDENTIFICATION AND VERIFICATION OF ACCIDENT VICTIMS. AfriResearch Platform.