Abstract

This research proposes a new insight into the iris recognition capture devices traditionally used to identify people. Now, we are using the same capture device to extract complementary information from periocular near-infrared iris images to detect the reduction of alertness conditions due to alcohol, drugs and sleepiness. The study focuses on determining the effect of external impairing influences on the Central Nervous System. The goal is to analyse how these factors impact eye/iris/pupil movement and if it is possible to classify these changes with a standard iris NIR device. This paper proposes a modified MobileNetV3, which focuses on fine textures per each RGB channel to classify subjects under the influences of alcohol/drugs/sleepiness into fit and unfit. Further, a new four-class database called ”FFD-NIR-Stream” was created. The results show that the MobileNetV3 classifier can detect the unfit condition from iris samples captured after alcohol/drug consumption with an accuracy of 91.3% and 99.1%, respectively. The sleepiness condition is the most challenging, with 72.4%. For two classes grouped on ”Fit/Unfit”, the model obtained an accuracy of 94.0% and 84.0%, respectively. This work is a step forward in iris biometric applications. © 2025 The Authors