CUST-Iris

Abstract

The CUST-Iris dataset was collected at Capital University of Science and Technology (CUST), Islamabad, Pakistan, for biometric iris recognition research. The dataset contains 2,880 iris images acquired from 720 unique irises (L+R) using the Crossmatch I-Scan 2 iris scanner. Images were captured in the near-infrared (NIR) spectrum with a resolution of 480 × 480 pixels. Each identity contributes images from both eyes, resulting in a diverse collection of iris samples suitable for verification, identification, segmentation, and federated learning research. The dataset is organized using a structured naming convention of the form Identity_Eye_Image, represented as: Identity_Eye_Image where: • Identity denotes the subject identifier. • Eye denotes the eye being captured (1 or 2). • Image denotes the image number for that eye. For example: • 1_1_1 represents the first image of Eye 1 from Identity 1. • 1_1_2 represents the second image of Eye 1 from Identity 1. • 1_2_1 represents the first image of Eye 2 from Identity 1. The CUST-Iris dataset was collected with explicit permission from both the participating subjects and the host institution, Capital University of Science and Technology (CUST), Islamabad, Pakistan. All data acquisition procedures were conducted in accordance with institutional ethical guidelines, and informed consent was obtained from all participants prior to imaging. The subjects were made aware that the collected iris images would be used for academic research purposes and may be shared with the research community in anonymized form. No personally identifiable information is included in the dataset, and all identities are represented using numerical labels to ensure privacy and compliance with ethical standards.