Low-Power Pedestrian Detection System on FPGA
Pedestrian detection is one of the key problems in the emerging self-driving car industry. In addition, the Histogram of Gradients (HOG) algorithm proved to provide good accuracy for pedestrian detection. Many research works focused on accelerating HOG algorithm on FPGA (Field-Programmable Gate Arra...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2019-11-01
|
Series: | Proceedings |
Subjects: | |
Online Access: | https://www.mdpi.com/2504-3900/31/1/35 |
id |
doaj-035bb0f648394c9ba7024685bda71d7e |
---|---|
record_format |
Article |
spelling |
doaj-035bb0f648394c9ba7024685bda71d7e2020-11-24T21:51:05ZengMDPI AGProceedings2504-39002019-11-013113510.3390/proceedings2019031035proceedings2019031035Low-Power Pedestrian Detection System on FPGAVinh Ngo0David Castells-Rufas1Arnau Casadevall2Marc Codina3Jordi Carrabina4Department of Microelectronics and Electronic Systems, School of Engineering, Autonomous University of Barcelona, 08193 Bellaterra, SpainDepartment of Microelectronics and Electronic Systems, School of Engineering, Autonomous University of Barcelona, 08193 Bellaterra, SpainDepartment of Microelectronics and Electronic Systems, School of Engineering, Autonomous University of Barcelona, 08193 Bellaterra, SpainDepartment of Microelectronics and Electronic Systems, School of Engineering, Autonomous University of Barcelona, 08193 Bellaterra, SpainDepartment of Microelectronics and Electronic Systems, School of Engineering, Autonomous University of Barcelona, 08193 Bellaterra, SpainPedestrian detection is one of the key problems in the emerging self-driving car industry. In addition, the Histogram of Gradients (HOG) algorithm proved to provide good accuracy for pedestrian detection. Many research works focused on accelerating HOG algorithm on FPGA (Field-Programmable Gate Array) due to its low-power and high-throughput characteristics. In this paper, we present an energy-efficient HOG-based implementation for pedestrian detection system on a low-cost FPGA system-on-chip platform. The hardware accelerator implements the HOG computation and the Support Vector Machine classifier, the rest of the algorithm is mapped to software in the embedded processor. The hardware runs at 50 Mhz (lower frequency than previous works), thus achieving the best pixels processed per clock and the lower power design.https://www.mdpi.com/2504-3900/31/1/35fpgahog extractorpedestrian detectionacceleratorlow power |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Vinh Ngo David Castells-Rufas Arnau Casadevall Marc Codina Jordi Carrabina |
spellingShingle |
Vinh Ngo David Castells-Rufas Arnau Casadevall Marc Codina Jordi Carrabina Low-Power Pedestrian Detection System on FPGA Proceedings fpga hog extractor pedestrian detection accelerator low power |
author_facet |
Vinh Ngo David Castells-Rufas Arnau Casadevall Marc Codina Jordi Carrabina |
author_sort |
Vinh Ngo |
title |
Low-Power Pedestrian Detection System on FPGA |
title_short |
Low-Power Pedestrian Detection System on FPGA |
title_full |
Low-Power Pedestrian Detection System on FPGA |
title_fullStr |
Low-Power Pedestrian Detection System on FPGA |
title_full_unstemmed |
Low-Power Pedestrian Detection System on FPGA |
title_sort |
low-power pedestrian detection system on fpga |
publisher |
MDPI AG |
series |
Proceedings |
issn |
2504-3900 |
publishDate |
2019-11-01 |
description |
Pedestrian detection is one of the key problems in the emerging self-driving car industry. In addition, the Histogram of Gradients (HOG) algorithm proved to provide good accuracy for pedestrian detection. Many research works focused on accelerating HOG algorithm on FPGA (Field-Programmable Gate Array) due to its low-power and high-throughput characteristics. In this paper, we present an energy-efficient HOG-based implementation for pedestrian detection system on a low-cost FPGA system-on-chip platform. The hardware accelerator implements the HOG computation and the Support Vector Machine classifier, the rest of the algorithm is mapped to software in the embedded processor. The hardware runs at 50 Mhz (lower frequency than previous works), thus achieving the best pixels processed per clock and the lower power design. |
topic |
fpga hog extractor pedestrian detection accelerator low power |
url |
https://www.mdpi.com/2504-3900/31/1/35 |
work_keys_str_mv |
AT vinhngo lowpowerpedestriandetectionsystemonfpga AT davidcastellsrufas lowpowerpedestriandetectionsystemonfpga AT arnaucasadevall lowpowerpedestriandetectionsystemonfpga AT marccodina lowpowerpedestriandetectionsystemonfpga AT jordicarrabina lowpowerpedestriandetectionsystemonfpga |
_version_ |
1725880557522386944 |