Back to search

West Virginia University

Deep-Learning-Based Generation of Synthetic Contactless Fingerphotos

Abstract

dc:description.abstract

The collection of biometric data is a labor-intensive, high-resource process that presents significant logistical, privacy, and cost barriers for researchers and developers. To address these challenges, the biometrics community has increasingly turned to generative models capable of producing synthetic datasets that reflect the statistical properties of real data. While substantial progress has been made in synthetic fingerprint generation for contact-based modalities, the contactless fingerphoto domain has remained largely underserved. This work presents a deep learning-based approach to synthetic contactless fingerphoto generation using a Stable Diffusion model guided by multimodal conditions (text and image). The dataset used for training was collected by the WVU Biometrics Lab using smartphones to captures fingerphotos in optimal and suboptimal conditions. The proposed framework generates synthetic contactless fingerphotos across five fingerprint pattern classes (Arch, Tented Arch, Left Loop, Right Loop, and Whorl) with structural fidelity enforced through Gabor ridge orientation maps that are guided using ControlNet to direct the diffusion process toward accurate ridge geometry and spatial frequency. A composite loss function incorporating perceptual, Gabor ridge, contrast, and FFT spectral components further constrains generation to preserve the meaningful structure of the contactless fingerprint domain. The resulting synthetic dataset is evaluated on fingerprint-specific image quality metrics, perceptual distributional fidelity, and matcher-based biometric utility, demonstrating the feasibility of large-scale synthetic contactless fingerphoto generation for use in training and evaluating contactless fingerprint recognition systems.

Degree

thesis:*
Name thesis:degree_name
MS
Level thesis:degree_level
Thesis
Discipline thesis:degree_discipline
Lane Department of Computer Science and Electrical Engineering

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Burton, Christopher Harry
Contributors dc:contributor
  • Jeremy Dawson
  • Prashnna Gyawali
  • Nima Karimian

Subjects

dc:subject × 4

Identifiers

dc:identifier.*
Identifier
<p><a href="https://orcid.org/0009-0004-7602-7741">https://orcid.org/0009-0004-7602-7741</a></p>
https://researchrepository.wvu.edu/etd/13242

Chain of custody

source
Harvested from
West Virginia University
Base URL
researchrepository.wvu.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Burton, Christopher Harry. Deep-Learning-Based Generation of Synthetic Contactless Fingerphotos. Thesis thesis, https://doi.org/10.33915/etd.13242