Deep Learning Techniques for User Authentication in Medical IoT Applications

View at publisher Summarise with VIRAT
Authors: Lithick Sanjaiyan S, Mohammed Ansar Ali M A, Kanagaviswesh G, Sugansingh Kumaran K S
Cited by: 0
References: 16

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The rapid integration of the Internet of Medical Things (IoMT) into healthcare systems has introduced new challenges in secure, contactless, and real-time user authentication. Conventional authentication mechanisms such as passwords and access cards are inefficient in medical environments where hygiene, speed, and hands-free interaction are critical. Face recognition has emerged as a viable alternative; however, recognition performance under unconstrained environments remains a challenge due to pose variation, illumination changes, occlusions, and background clutter. This work presents a comprehensive comparative analysis of five deep learning and hybrid face recognition models evaluated as an authentication framework for IoMT devices. Experiments conducted on the Labeled Faces in the Wild (LFW) dataset demonstrate that metric learning approaches significantly improve discriminative embedding quality. The ArcFace backbone improves authentication accuracy from 39.26% to 76.44%, while ArcFace embeddings combined with SVM achieve the highest accuracy of 78.06%. Transformer-based fusion underperforms due to limited dataset scale. The study highlights that embedding geometry and margin-based optimization are more critical than architectural complexity for reliable IoMT user authentication.

Fields of Study & Subjects

Face recognition and analysis Biometric Identification and Security Face and Expression Recognition

Publication details

Journal / source
International Conference on Computing, Communication, Control and Cyber-Physical Systems (I5CPS)
Pages
1–5
Publication date
26 February 2026
Language
EN