A neuro-inspired visual tracking method based on programmable system-on-chip platform

Shufan Yang, KongFatt Wong-Lin, James Andrew, Terrence Mak, TM McGinnity

Research output: Contribution to journalArticle

Abstract

Using programmable system-on-chip to implement computer vision functions poses many challenges due to highly constrained resources in cost, size and power consumption. In this work, we propose a new neuro-inspired image processing model and implemented it on a system-on-chip Xilinx Z702c board. With the attractor neural network model to store the object’s contour information, we eliminate the computationally expensive steps in the curve evolution re-initialization at every new iteration or frame. Our experimental results demonstrate that this integrated approach achieves accurate and robust object tracking, when they are partially or completely occluded in the scenes. Importantly, the system is able to process 640 by 480 videos in real-time stream with 30 frames per second using only one low-power Xilinx Zynq-7000 system-on-chip platform. This proof-of-concept work has demonstrated the advantage of incorporating neuro-inspired features in solving image processing problems during occlusion.
LanguageEnglish
Pages2697-2708
JournalNeural Computing and Applications
Volume30
Issue number9
Early online date20 Jan 2017
DOIs
Publication statusPublished - Nov 2018

Fingerprint

Image processing
Computer vision
Electric power utilization
Neural networks
System-on-chip
Costs

Keywords

  • Visual object tracking
  • mean-shift
  • level set
  • attractor neural network model
  • occlusion
  • system-on-chip

Cite this

Yang, Shufan ; Wong-Lin, KongFatt ; Andrew, James ; Mak, Terrence ; McGinnity, TM. / A neuro-inspired visual tracking method based on programmable system-on-chip platform. 2018 ; Vol. 30, No. 9. pp. 2697-2708.
@article{fe521b1304c141908fd5eaad3d4ed405,
title = "A neuro-inspired visual tracking method based on programmable system-on-chip platform",
abstract = "Using programmable system-on-chip to implement computer vision functions poses many challenges due to highly constrained resources in cost, size and power consumption. In this work, we propose a new neuro-inspired image processing model and implemented it on a system-on-chip Xilinx Z702c board. With the attractor neural network model to store the object’s contour information, we eliminate the computationally expensive steps in the curve evolution re-initialization at every new iteration or frame. Our experimental results demonstrate that this integrated approach achieves accurate and robust object tracking, when they are partially or completely occluded in the scenes. Importantly, the system is able to process 640 by 480 videos in real-time stream with 30 frames per second using only one low-power Xilinx Zynq-7000 system-on-chip platform. This proof-of-concept work has demonstrated the advantage of incorporating neuro-inspired features in solving image processing problems during occlusion.",
keywords = "Visual object tracking, mean-shift, level set, attractor neural network model, occlusion, system-on-chip",
author = "Shufan Yang and KongFatt Wong-Lin and James Andrew and Terrence Mak and TM McGinnity",
year = "2018",
month = "11",
doi = "10.1007/s00521-017-2847-5",
language = "English",
volume = "30",
pages = "2697--2708",
number = "9",

}

A neuro-inspired visual tracking method based on programmable system-on-chip platform. / Yang, Shufan; Wong-Lin, KongFatt; Andrew, James; Mak, Terrence; McGinnity, TM.

Vol. 30, No. 9, 11.2018, p. 2697-2708.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A neuro-inspired visual tracking method based on programmable system-on-chip platform

AU - Yang, Shufan

AU - Wong-Lin, KongFatt

AU - Andrew, James

AU - Mak, Terrence

AU - McGinnity, TM

PY - 2018/11

Y1 - 2018/11

N2 - Using programmable system-on-chip to implement computer vision functions poses many challenges due to highly constrained resources in cost, size and power consumption. In this work, we propose a new neuro-inspired image processing model and implemented it on a system-on-chip Xilinx Z702c board. With the attractor neural network model to store the object’s contour information, we eliminate the computationally expensive steps in the curve evolution re-initialization at every new iteration or frame. Our experimental results demonstrate that this integrated approach achieves accurate and robust object tracking, when they are partially or completely occluded in the scenes. Importantly, the system is able to process 640 by 480 videos in real-time stream with 30 frames per second using only one low-power Xilinx Zynq-7000 system-on-chip platform. This proof-of-concept work has demonstrated the advantage of incorporating neuro-inspired features in solving image processing problems during occlusion.

AB - Using programmable system-on-chip to implement computer vision functions poses many challenges due to highly constrained resources in cost, size and power consumption. In this work, we propose a new neuro-inspired image processing model and implemented it on a system-on-chip Xilinx Z702c board. With the attractor neural network model to store the object’s contour information, we eliminate the computationally expensive steps in the curve evolution re-initialization at every new iteration or frame. Our experimental results demonstrate that this integrated approach achieves accurate and robust object tracking, when they are partially or completely occluded in the scenes. Importantly, the system is able to process 640 by 480 videos in real-time stream with 30 frames per second using only one low-power Xilinx Zynq-7000 system-on-chip platform. This proof-of-concept work has demonstrated the advantage of incorporating neuro-inspired features in solving image processing problems during occlusion.

KW - Visual object tracking

KW - mean-shift

KW - level set

KW - attractor neural network model

KW - occlusion

KW - system-on-chip

U2 - 10.1007/s00521-017-2847-5

DO - 10.1007/s00521-017-2847-5

M3 - Article

VL - 30

SP - 2697

EP - 2708

IS - 9

ER -