Loading

Real-Time Iris Detection and Recognition System Using You Only Look Once Version 8
Parthasarathy C1, Priscilla Rachel G2, Rachel Sherin J3

1Parthasarathy C, Department of Computer Science and Engineering Rajalakshmi Engineering College Chennai (Tamil Nadu), India.

2Priscilla Rachel G, Department of Computer Science and Engineering Rajalakshmi Engineering College Chennai (Tamil Nadu), India.

3Rachel Sherin J, Department of Computer Science and Engineering Rajalakshmi Engineering College Chennai (Tamil Nadu), India.  

Manuscript received on 30 November 2024 | First Revised Manuscript received on 11 December 2024 | Second Revised Manuscript received on 25 January 2025 | Manuscript Accepted on 15 February 2025 | Manuscript published on 28 February 2025 | PP: 13-17 | Volume-12 Issue-2, February 2025 | Retrieval Number: 100.1/ijies.L109611121224 | DOI: 10.35940/ijies.L1096.12020225

Open Access | Editorial and Publishing Policies | Cite | Zenodo | OJS | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: The model used in real-time object detection, which offers high speed and accuracy by processing images in a single pass, is the You Only Look Once (YOLO) model. This project primarily focuses on the application of YOLOv8, or You Only Look Once version 8, for iris detection and recognition in biometric systems, with a focus on high security and accuracy. To improve the model’s performance under various lighting conditions, it was trained on customised datasets. To improve the model’s generalisation, advanced image augmentation techniques, such as flips, rotations, and brightness adjustments, were employed. The model yielded an average precision of 95% on the validation set, which was trained using the PyTorch framework with optimised hyperparameters, demonstrating the effectiveness of YOLOv8 in real-time iris recognition and detection.

Keywords: Iris Recognition System, Image Augmentation, Pytorch Framework, Hyper Parameter, Generalization.
Scope of the Article: Computer Science and Applications