KIT | KIT-Bibliothek | Impressum | Datenschutz

Embedded Face Recognition for Personalized Services in the Assistive Robotics

Walter, Iris ORCID iD icon 1; Ney, Jonas; Hotfilter, Tim ORCID iD icon 1; Rybalkin, Vladimir; Hoefer, Julian ORCID iD icon 1; Wehn, Norbert; Becker, Jürgen 1
1 Karlsruher Institut für Technologie (KIT)


Recently, the field of assistive robotics has drawn much attention in the health care sector. In combination with modern machine learning-supported person recognition systems, they can deliver highly personalized services. However, common algorithms for person recognition such as convolutional neural networks (CNNs) consume high amounts of power and show low energy efficiency when executed on general-purpose computing platforms.
In this paper, we present our hardware architecture and field programmable gate array (FPGA) accelerator to enable on-device person recognition in the context of assistive robotics. Therefore, we optimize a neural network based on the SqueezeNet topology and implement it on an FPGA for a high degree of flexibility and reconfigurability. By pruning redundant filters and quantization of weights and activations, we are able to find a well-fitting neural network that achieves a high identification accuracy of 84%. On a Xilinx Zynq Ultra96v2, we achieve a power consumption of 4.8 W, a latency of 31 ms and an efficiency of 6.738 FPS/W. Our results outperform the latency by 1.6x compared to recent person recognition systems in assistive robots and energy efficiency by 1.7x for embedded face recognition, respectively.

Postprint §
DOI: 10.5445/IR/1000143166
Veröffentlicht am 02.01.2023
DOI: 10.1007/978-3-030-93736-2_26
Zitationen: 5
Zitationen: 7
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Technik der Informationsverarbeitung (ITIV)
Publikationstyp Proceedingsbeitrag
Publikationsjahr 2022
Sprache Englisch
Identifikator ISBN: 978-3-030-93736-2
ISSN: 1865-0929, 1865-0937
KITopen-ID: 1000143166
Erschienen in Machine Learning and Principles and Practice of Knowledge Discovery in Databases – International Workshops of ECML PKDD 2021, Virtual Event, September 13-17, 2021, Proceedings, Part I. Ed.: M. Kamp
Veranstaltung International Workshops of European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2021), Online, 13.09.2021 – 17.09.2021
Verlag Springer International Publishing
Seiten 339–350
Serie Communications in Computer and Information Science
Vorab online veröffentlicht am 01.01.2022
Schlagwörter Ambient assisted living, Assistive robotics, Convolutional neural networks, Face recognition, Field programmable, gate array, Quantization
Nachgewiesen in Scopus
Relationen in KITopen
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page