- Client
- Criminal Record Division of the Criminal Investigation Department, Sri Lanka Police
- Development
- Theekshana, building on the UCSC Software Development Unit’s research and engineering capacity
- Project began
- 2010
- Official launch
- February 2014
- Initial funding
- Information and Communication Technology Agency of Sri Lanka
- Recognition
- e-Swabhimani 2014; Manthan Award recognition in 2015
The challenge
In November 2010, Harsha Wijayawardhana was asked to undertake the development of an Automated Fingerprint Identification System for Sri Lanka Police. The Criminal Record Division faced a substantial backlog of fingerprint records accumulated over many years. Imported commercial systems would have imposed very high acquisition and continuing costs, while still requiring adaptation to local operational practices.
Harsha approached the assignment as both an engineering problem and a national-capacity problem: Sri Lanka needed software that could be researched, built, understood, maintained and progressively improved by a Sri Lankan team.
Research before implementation
The work took roughly two years of research, algorithm development, testing and close engagement with fingerprint examiners. Under Harsha’s technical leadership, the team developed matching and parallel-processing algorithms together with examiner tools for cleaning disturbed latent prints, cropping images and marking minutiae.
The resulting platform was designed to support ten-print and latent-print workflows, including ten-print-to-ten-print searching and the comparison of crime-scene fragments against the growing national collection. Image processing and matching components were combined with database, web and examiner-facing software rather than treated as a single isolated algorithm.
Digitisation and operational preparation
By 2013, the Criminal Record Division was digitising its fingerprint holdings in preparation for full operational use. Harsha’s project record notes that approximately 900,000 ten-print records had entered the digital collection. The system combined automated candidate searching with the professional judgement of trained fingerprint examiners, preserving the human verification required in forensic work.
The 2014 launch
AFIS was officially launched in February 2014. The launch marked the transition from a difficult research and development programme into a national operational system. Contemporary reporting described the system as a UCSC–ICTA fingerprint innovation, highlighted its locally developed algorithms, and reported substantial savings compared with purchasing an overseas platform.
For Harsha, the launch represented the culmination of sustained work by developers, fingerprint specialists, Police leadership, Theekshana, UCSC and ICTA. It also demonstrated the value of maintaining an applied software-research team capable of responding to problems of national importance.
Expansion and recognition
The system received the e-Swabhimani award for e-Government and Institutions in 2014 and gained regional recognition through the Manthan Awards in 2015. In December 2015, the service was expanded to 42 Police divisions, bringing fingerprint-search capability closer to investigators around the country.
AFIS remains part of a wider body of national software led by Harsha through the UCSC Software Development Unit and Theekshana: election systems, civil-registration digitisation, university-selection software, agricultural information systems and other public-sector platforms. Among them, AFIS stands out for the depth of its applied research, its forensic consequences and its demonstration that advanced biometric software could be engineered locally.
- Theekshana project record — project period, client, funding and award.
- Daily FT, 19 March 2014 — contemporary report on the launch, technical work and savings.
- Daily FT, 30 December 2015 — expansion to 42 Police divisions.
- Mint, 20 January 2016 — independent feature on development, use and Manthan recognition.
- Preserved AFIS media index — further reporting and photographs.