Accelerating computer vision algorithms on heterogeneous edge computing platforms
Heterogeneity has become the cornerstone of modern embedded System-on-Chips (SoCs), used in latest smart-phones and edge computing platforms to achieve high performance under tight power budgets. Alongside the general purpose multi-core CPUs, such SoCs typically integrate several specialized process...
Saved in:
Main Authors: | , , , |
---|---|
Other Authors: | |
Format: | Conference or Workshop Item |
Language: | English |
Published: |
2021
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/147717 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-147717 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1477172021-04-20T07:27:32Z Accelerating computer vision algorithms on heterogeneous edge computing platforms Prakash, Alok Ramakrishnan, Nirmala Garg, Kratika Srikanthan, Thambipillai School of Computer Science and Engineering 2020 IEEE Workhop on Signal Processing Systems (SiPS) Engineering::Computer science and engineering KLT Computer Vision Heterogeneity has become the cornerstone of modern embedded System-on-Chips (SoCs), used in latest smart-phones and edge computing platforms to achieve high performance under tight power budgets. Alongside the general purpose multi-core CPUs, such SoCs typically integrate several specialized processing elements such as GPU and DSP to ensure power-efficient execution of specific workloads. In the past, many computer vision algorithms and their kernels have been shown to benefit from execution on GPUs, both in terms of performance and power consumption. Existing work has also demonstrated the benefit of accelerating them simultaneously on multi-core CPUs and discrete as well as integrated GPUs found in PCs and workstations. Recently, authors have also focused on accelerating such applications on heterogeneous embedded SoCs with integrated CPU and GPU. In this paper, we first present an extensive literature review of such efforts and highlight their strengths and limitations. Next, we use the latest state-of-the-art edge computing platform, Odroid-N2, and the older Odroid-XU3 platform, both of which use heterogeneous embedded SoCs, to explore the acceleration of the convolution kernel with different filter sizes and its impact on the KLT tracking algorithm. Lastly, we discuss the challenges and opportunities in leveraging such SoCs for computer vision and other AI applications. 2021-04-20T07:27:32Z 2021-04-20T07:27:32Z 2020 Conference Paper Prakash, A., Ramakrishnan, N., Garg, K. & Srikanthan, T. (2020). Accelerating computer vision algorithms on heterogeneous edge computing platforms. 2020 IEEE Workhop on Signal Processing Systems (SiPS), 2020-October, 1-6. https://dx.doi.org/10.1109/SiPS50750.2020.9195221 9781728180991 https://hdl.handle.net/10356/147717 10.1109/SiPS50750.2020.9195221 2-s2.0-85096776323 2020-October 1 6 en NRF TUMCREATE © 2020 Institute of Electrical and Electronics Engineers (IEEE). All rights reserved. |
institution |
Nanyang Technological University |
building |
NTU Library |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
NTU Library |
collection |
DR-NTU |
language |
English |
topic |
Engineering::Computer science and engineering KLT Computer Vision |
spellingShingle |
Engineering::Computer science and engineering KLT Computer Vision Prakash, Alok Ramakrishnan, Nirmala Garg, Kratika Srikanthan, Thambipillai Accelerating computer vision algorithms on heterogeneous edge computing platforms |
description |
Heterogeneity has become the cornerstone of modern embedded System-on-Chips (SoCs), used in latest smart-phones and edge computing platforms to achieve high performance under tight power budgets. Alongside the general purpose multi-core CPUs, such SoCs typically integrate several specialized processing elements such as GPU and DSP to ensure power-efficient execution of specific workloads. In the past, many computer vision algorithms and their kernels have been shown to benefit from execution on GPUs, both in terms of performance and power consumption. Existing work has also demonstrated the benefit of accelerating them simultaneously on multi-core CPUs and discrete as well as integrated GPUs found in PCs and workstations. Recently, authors have also focused on accelerating such applications on heterogeneous embedded SoCs with integrated CPU and GPU. In this paper, we first present an extensive literature review of such efforts and highlight their strengths and limitations. Next, we use the latest state-of-the-art edge computing platform, Odroid-N2, and the older Odroid-XU3 platform, both of which use heterogeneous embedded SoCs, to explore the acceleration of the convolution kernel with different filter sizes and its impact on the KLT tracking algorithm. Lastly, we discuss the challenges and opportunities in leveraging such SoCs for computer vision and other AI applications. |
author2 |
School of Computer Science and Engineering |
author_facet |
School of Computer Science and Engineering Prakash, Alok Ramakrishnan, Nirmala Garg, Kratika Srikanthan, Thambipillai |
format |
Conference or Workshop Item |
author |
Prakash, Alok Ramakrishnan, Nirmala Garg, Kratika Srikanthan, Thambipillai |
author_sort |
Prakash, Alok |
title |
Accelerating computer vision algorithms on heterogeneous edge computing platforms |
title_short |
Accelerating computer vision algorithms on heterogeneous edge computing platforms |
title_full |
Accelerating computer vision algorithms on heterogeneous edge computing platforms |
title_fullStr |
Accelerating computer vision algorithms on heterogeneous edge computing platforms |
title_full_unstemmed |
Accelerating computer vision algorithms on heterogeneous edge computing platforms |
title_sort |
accelerating computer vision algorithms on heterogeneous edge computing platforms |
publishDate |
2021 |
url |
https://hdl.handle.net/10356/147717 |
_version_ |
1698713716971798528 |