@Article{cmes.2026.081429,
AUTHOR = {Karthick Raghunath K. M., Manjula V., Mahesh T. R., Surbhi B. Khan, Ahmed Alyahya, Shakila Basheer},
TITLE = {Quantum-Enhanced Transparent Contour-Integral Deep Learning for Fair and Explainable Medical Biometric Authentication},
JOURNAL = {Computer Modeling in Engineering \& Sciences},
VOLUME = {148},
YEAR = {2026},
NUMBER = {1},
PAGES = {0--0},
URL = {http://www.techscience.com/CMES/v148n1/68186},
ISSN = {1526-1506},
ABSTRACT = {In general, medical biometric datasets, with the essential unique behavioral and physical traits for personalized healthcare, shape the patient identification process, but the tendency towards transparency and fairness is still far away. Most of the existing methods fail to integrate the latest mathematical techniques rigorously with the deep learning models, which eventually makes such models undesirable due to their lack of interpretability and potential bias. As such, in this study, a novel Contour Integrated Transparent Augmented Deep Learning (CITADL) methodology is introduced to bridge this gap. In this study, a structured framework, namely CITADL, combines contour-based mathematical feature transformation with a deep neural network architecture empowered with eXplainable Artificial Intelligence (XAI) modules. By using numerically grounded contour integration techniques within a deep learning module, CITADL detects temporal and spatial patterns in biometric signals based on adaptive fairness regularization and transparency layers, which allow it to explain the decision-making process. The framework is further supplemented through a quantum feature regeneration module that is designed to encode contour-integrated biometric representations in a Variational Quantum Circuit (VQC) and generate enhanced nonlinear correlation modelling. The resulting measured qubits features are decoded and harmoniously injected into the classical deep-learning pipeline, thus retaining the transparency, fairness and explainability. The evaluations of preliminary studies show 97% relative improvement in predictive accuracy and fairness metrics compared to selected baseline models in clinical biometric applications, thus contributing to the development of a reliable and transparent methodology for ethical augmentation of medical biometric data.},
DOI = {10.32604/cmes.2026.081429}
}