Use this URL to cite or link to this record in EThOS:
Title: A systematic approach to fingerprint identification via source probabilities
Author: Pillin, Etienne J. A.
ISNI:       0000 0004 8498 3832
Awarding Body: University of Leicester
Current Institution: University of Leicester
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
This research project was carried out under the INTREPID Forensics programme, a doctoral program involving 10 Ph.D. students in various fields applied to Forensic Science, and funded by the European Commission. The purpose of this project was to produce innovative methods of pattern recognition for fingerprint ridge lines in order to improve the reliability and the amenability of automatic fingerprint identification to the court. This research provides a preliminary but systematic and necessary approach to achieve this. First of all, the premise, software, and methodology for a data collection were developed for the purpose of a ground-truth database suitable for research and the training of identification algorithms. Two novel mathematical formulations of the fingerprint identification problem were made: source determination and source assessment. The latter provides a basis for the computation of source probabilities, namely the probability for two finger impressions to come from the same source, which is not considered sound. Despite current consensus, this thesis has established a new approach that proves that this can in fact be done in a mathematically justified manner. Finally, this research culminated with the development of feature detection algorithms that proceed by fitting a section of a fingerprint image by a function which locally models the ridge line accurately, and which demonstrated promising results. The fitting methods used rely on optimisation algorithms known as Estimation of Distribution Algorithms (EDAs), which have been generalised to the context of mixed-discrete optimisation, and implemented and applied to fingerprint feature detection.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
Keywords: Thesis