A task-oriented vision system

Yang Xiao, Kevin Irick, John Morgan Sampson, Vijaykrishnan Narayanan, Chuanjun Zhang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Recently, biologically inspired vision systems have been the focus of intense research effort to emulate the high energy-efficiency, performance and robustness of mammalian vision systems. However, previous vision accelerators have only focused on speeding up computationally intense portions of the system without exploiting effects seen in the human brain that demonstrate the task influence in the vision mechanism. In this paper, we propose a task-oriented two-level vision system which is composed of Saliency and SURF. To the best of our knowledge, our design is the first embedded system that utilizes task influence in the computation of visual attention and recognition. As a result, we show that the new system can achieve at most 12.75% accuracy improvement while saving 25% computation work.

Original languageEnglish (US)
Title of host publicationGLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI
PublisherAssociation for Computing Machinery
Pages181-186
Number of pages6
ISBN (Print)9781450328166
DOIs
StatePublished - Jan 1 2014
Event24th Great Lakes Symposium on VLSI, GLSVLSI 2014 - Houston, TX, United States
Duration: May 21 2014May 23 2014

Publication series

NameProceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI

Other

Other24th Great Lakes Symposium on VLSI, GLSVLSI 2014
CountryUnited States
CityHouston, TX
Period5/21/145/23/14

Fingerprint

Embedded systems
Particle accelerators
Energy efficiency
Brain

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Xiao, Y., Irick, K., Sampson, J. M., Narayanan, V., & Zhang, C. (2014). A task-oriented vision system. In GLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI (pp. 181-186). (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI). Association for Computing Machinery. https://doi.org/10.1145/2591513.2591602
Xiao, Yang ; Irick, Kevin ; Sampson, John Morgan ; Narayanan, Vijaykrishnan ; Zhang, Chuanjun. / A task-oriented vision system. GLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI. Association for Computing Machinery, 2014. pp. 181-186 (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI).
@inproceedings{c16229181a3c4debb890eecc66ff448d,
title = "A task-oriented vision system",
abstract = "Recently, biologically inspired vision systems have been the focus of intense research effort to emulate the high energy-efficiency, performance and robustness of mammalian vision systems. However, previous vision accelerators have only focused on speeding up computationally intense portions of the system without exploiting effects seen in the human brain that demonstrate the task influence in the vision mechanism. In this paper, we propose a task-oriented two-level vision system which is composed of Saliency and SURF. To the best of our knowledge, our design is the first embedded system that utilizes task influence in the computation of visual attention and recognition. As a result, we show that the new system can achieve at most 12.75{\%} accuracy improvement while saving 25{\%} computation work.",
author = "Yang Xiao and Kevin Irick and Sampson, {John Morgan} and Vijaykrishnan Narayanan and Chuanjun Zhang",
year = "2014",
month = "1",
day = "1",
doi = "10.1145/2591513.2591602",
language = "English (US)",
isbn = "9781450328166",
series = "Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI",
publisher = "Association for Computing Machinery",
pages = "181--186",
booktitle = "GLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI",

}

Xiao, Y, Irick, K, Sampson, JM, Narayanan, V & Zhang, C 2014, A task-oriented vision system. in GLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI. Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI, Association for Computing Machinery, pp. 181-186, 24th Great Lakes Symposium on VLSI, GLSVLSI 2014, Houston, TX, United States, 5/21/14. https://doi.org/10.1145/2591513.2591602

A task-oriented vision system. / Xiao, Yang; Irick, Kevin; Sampson, John Morgan; Narayanan, Vijaykrishnan; Zhang, Chuanjun.

GLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI. Association for Computing Machinery, 2014. p. 181-186 (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - A task-oriented vision system

AU - Xiao, Yang

AU - Irick, Kevin

AU - Sampson, John Morgan

AU - Narayanan, Vijaykrishnan

AU - Zhang, Chuanjun

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Recently, biologically inspired vision systems have been the focus of intense research effort to emulate the high energy-efficiency, performance and robustness of mammalian vision systems. However, previous vision accelerators have only focused on speeding up computationally intense portions of the system without exploiting effects seen in the human brain that demonstrate the task influence in the vision mechanism. In this paper, we propose a task-oriented two-level vision system which is composed of Saliency and SURF. To the best of our knowledge, our design is the first embedded system that utilizes task influence in the computation of visual attention and recognition. As a result, we show that the new system can achieve at most 12.75% accuracy improvement while saving 25% computation work.

AB - Recently, biologically inspired vision systems have been the focus of intense research effort to emulate the high energy-efficiency, performance and robustness of mammalian vision systems. However, previous vision accelerators have only focused on speeding up computationally intense portions of the system without exploiting effects seen in the human brain that demonstrate the task influence in the vision mechanism. In this paper, we propose a task-oriented two-level vision system which is composed of Saliency and SURF. To the best of our knowledge, our design is the first embedded system that utilizes task influence in the computation of visual attention and recognition. As a result, we show that the new system can achieve at most 12.75% accuracy improvement while saving 25% computation work.

UR - http://www.scopus.com/inward/record.url?scp=84902588314&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84902588314&partnerID=8YFLogxK

U2 - 10.1145/2591513.2591602

DO - 10.1145/2591513.2591602

M3 - Conference contribution

AN - SCOPUS:84902588314

SN - 9781450328166

T3 - Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI

SP - 181

EP - 186

BT - GLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI

PB - Association for Computing Machinery

ER -

Xiao Y, Irick K, Sampson JM, Narayanan V, Zhang C. A task-oriented vision system. In GLSVLSI 2014 - Proceedings of the 2014 Great Lakes Symposium on VLSI. Association for Computing Machinery. 2014. p. 181-186. (Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI). https://doi.org/10.1145/2591513.2591602