Face detection with YOLO on edge.

Ali-Gombe, Adamu; Elyan, Eyad; Moreno-García, Carlos Francisco; Zwiegelaar, Johan

Abstract

Significant progress has been achieved in objects detection applications such as Face Detection. This mainly due to the latest development in deep learning-based approaches and especially in the computer vision domain. However, deploying deep-learning methods require huge computational power such as graphical processing units. These computational requirements make using such methods unsuitable for deployment on platforms with limited resources, such as edge devices. In this paper, we present an experimental framework to reduce the model’s size systematically, aiming at obtaining a small-size model suitable for deployment in a resource-limited environment. This was achieved by systematic layer removal and filter resizing. Extensive experiments were carried out using the “You Only Look Once” model (YOLO v3-tiny). For evaluation purposes, we used two public datasets to assess the impact of the model’s size reduction on a common computer vision task such as face detection. Results show clearly that, a significant reduction in the model’s size, has a very marginal impact on the overall model’s performance. These results open new directions towards further investigation and research to accelerate the use of deep learning models on edge-devices.

Citation

ALI-GOMBE, A., ELYAN, E., MORENO-GARCIA, C.F. and ZWIEGELAAR, J. 2021. Face detection with YOLO on edge. In Iliadis, L., Macintyre, J., Jayne, C. and Pimenidis, E. (eds.). Proceedings of the 22nd Enginering applications of neural networks conference (EANN2021), 25-27 June 2021, Halkidiki, Greece. Proceedings of the International Neural Networks Society (INNS), 3. Cham: Springer [online], pages 284-292. Available from: https://doi.org/10.1007/978-3-030-80568-5_24

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 22nd Enginering applications of neural networks conference (EANN2021)
• Start Date: Jun 25, 2021
• End Date: Jun 27, 2021
• Acceptance Date: Apr 7, 2021
• Online Publication Date: Jul 1, 2021
• Publication Date: Dec 31, 2021
• Deposit Date: Jun 25, 2021
• Publicly Available Date: Jul 2, 2022
• Publisher: Springer
• Peer Reviewed: Peer Reviewed
• Pages: 284-292
• Series Title: Proceedings of the International Neural Networks Society (INNS)
• Series Number: 3
• Series ISSN: 2661-8141
• Book Title: Proceedings of the 22nd Enginering applications of neural networks conference (EANN2021)
• ISBN: 9783030805678
• DOI: https://doi.org/10.1007/978-3-030-80568-5_24
• Keywords: Deep learning; YOLO; Face detection
• Public URL: https://rgu-repository.worktribe.com/output/1369880