SwePub
Sök i LIBRIS databas

  Extended search

id:"swepub:oai:gup.ub.gu.se/320431"
 

Search: id:"swepub:oai:gup.ub.gu.se/320431" > Real-time semantic ...

  • 1 of 1
  • Previous record
  • Next record
  •    To hitlist

Real-time semantic segmentation on FPGAs for autonomous vehicles with hls4ml

Ghielmetti, N. (author)
Organisation européenne pour la recherche nucléaire (CERN),European Organization for Nuclear Research (CERN),Politecnico di Milano,Polytechnic University of Milan
Loncar, V. (author)
Univerzitet u Beogradu,University of Belgrade,Organisation européenne pour la recherche nucléaire (CERN),European Organization for Nuclear Research (CERN)
Pierini, M. (author)
Organisation européenne pour la recherche nucléaire (CERN),European Organization for Nuclear Research (CERN)
show more...
Roed, M. (author)
University Of Oxford,Organisation européenne pour la recherche nucléaire (CERN),European Organization for Nuclear Research (CERN)
Summers, S. (author)
Organisation européenne pour la recherche nucléaire (CERN),European Organization for Nuclear Research (CERN)
Aarrestad, T. (author)
Eidgenössische Technische Hochschule Zürich (ETH),Swiss Federal Institute of Technology in Zürich (ETH)
Petersson, Christoffer, 1979 (author)
Chalmers tekniska högskola,Chalmers University of Technology
Linander, Hampus, 1985 (author)
Gothenburg University,Göteborgs universitet,Institutionen för fysik (GU),Department of Physics (GU),University of Gothenburg
Ngadiuba, J. (author)
Lin, K. L. (author)
University of Washington,Amazon
Harris, P. (author)
Massachusetts Institute of Technology (MIT)
show less...
 (creator_code:org_t)
2022-11-04
2022
English.
In: Machine Learning - Science and Technology. - : IOP Publishing. - 2632-2153. ; 3:4
  • Journal article (peer-reviewed)
Abstract Subject headings
Close  
  • In this paper, we investigate how field programmable gate arrays can serve as hardware accelerators for real-time semantic segmentation tasks relevant for autonomous driving. Considering compressed versions of the ENet convolutional neural network architecture, we demonstrate a fully-on-chip deployment with a latency of 4.9 ms per image, using less than 30% of the available resources on a Xilinx ZCU102 evaluation board. The latency is reduced to 3 ms per image when increasing the batch size to ten, corresponding to the use case where the autonomous vehicle receives inputs from multiple cameras simultaneously. We show, through aggressive filter reduction and heterogeneous quantization-aware training, and an optimized implementation of convolutional layers, that the power consumption and resource utilization can be significantly reduced while maintaining accuracy on the Cityscapes dataset.

Subject headings

NATURVETENSKAP  -- Data- och informationsvetenskap -- Datavetenskap (hsv//swe)
NATURAL SCIENCES  -- Computer and Information Sciences -- Computer Sciences (hsv//eng)
NATURVETENSKAP  -- Data- och informationsvetenskap -- Datorteknik (hsv//swe)
NATURAL SCIENCES  -- Computer and Information Sciences -- Computer Engineering (hsv//eng)
TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Kommunikationssystem (hsv//swe)
ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Communication Systems (hsv//eng)
NATURVETENSKAP  -- Data- och informationsvetenskap -- Datorseende och robotik (hsv//swe)
NATURAL SCIENCES  -- Computer and Information Sciences -- Computer Vision and Robotics (hsv//eng)

Keyword

FPGA
computer vision
deep learning
hls4ml
machine learning
autonomous vehicles
semantic segmentation
Computer Science
Science & Technology - Other Topics
machine learning

Publication and Content Type

ref (subject category)
art (subject category)

Find in a library

To the university's database

  • 1 of 1
  • Previous record
  • Next record
  •    To hitlist

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.

 
pil uppåt Close

Copy and save the link in order to return to this view