Toward a Generic Diver-Following Algorithm

Balancing Robustness and Efficiency in Deep Visual Detection

Md Jahidul Islam, Michael Fulton, Junaed Sattar

Research output: Contribution to journalArticle

Abstract

This letter explores the design and development of a class of robust diver detection algorithms for autonomous diver-following applications. By considering the operational challenges for underwater visual tracking in diverse real-world settings, we formulate a set of desired features of a generic diver-following algorithm. We attempt to accommodate these features and maximize general tracking performance by exploiting the state-of-the-art deep object detection models. We fine tune the building blocks of these models with a goal of balancing the tradeoff between robustness and efficiency in an on-board setting under real-time constraints. Subsequently, we design an architecturally simple convolutional neural network based diver detection model that is much faster than the state-of-the-art deep models yet provides comparable detection performances. In addition, we validate the performance and effectiveness of the proposed model through a number of diver-following experiments in closed-water and open-water environments.

Original languageEnglish (US)
Article number8543168
Pages (from-to)113-120
Number of pages8
JournalIEEE Robotics and Automation Letters
Volume4
Issue number1
DOIs
StatePublished - Jan 1 2019

Fingerprint

Balancing
Robustness
Water
Visual Tracking
Model
Object Detection
Building Blocks
Trade-offs
Maximise
Vision
Neural Networks
Neural networks
Real-time
Closed
Experiment
Experiments
Design

Keywords

  • Human detection and tracking
  • field robots
  • marine robotics

Cite this

Toward a Generic Diver-Following Algorithm : Balancing Robustness and Efficiency in Deep Visual Detection. / Islam, Md Jahidul; Fulton, Michael; Sattar, Junaed.

In: IEEE Robotics and Automation Letters, Vol. 4, No. 1, 8543168, 01.01.2019, p. 113-120.

Research output: Contribution to journalArticle

@article{417ed23357994bad95a9d31fbed8f363,
title = "Toward a Generic Diver-Following Algorithm: Balancing Robustness and Efficiency in Deep Visual Detection",
abstract = "This letter explores the design and development of a class of robust diver detection algorithms for autonomous diver-following applications. By considering the operational challenges for underwater visual tracking in diverse real-world settings, we formulate a set of desired features of a generic diver-following algorithm. We attempt to accommodate these features and maximize general tracking performance by exploiting the state-of-the-art deep object detection models. We fine tune the building blocks of these models with a goal of balancing the tradeoff between robustness and efficiency in an on-board setting under real-time constraints. Subsequently, we design an architecturally simple convolutional neural network based diver detection model that is much faster than the state-of-the-art deep models yet provides comparable detection performances. In addition, we validate the performance and effectiveness of the proposed model through a number of diver-following experiments in closed-water and open-water environments.",
keywords = "Human detection and tracking, field robots, marine robotics",
author = "Islam, {Md Jahidul} and Michael Fulton and Junaed Sattar",
year = "2019",
month = "1",
day = "1",
doi = "10.1109/LRA.2018.2882856",
language = "English (US)",
volume = "4",
pages = "113--120",
journal = "IEEE Robotics and Automation Letters",
issn = "2377-3766",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "1",

}

TY - JOUR

T1 - Toward a Generic Diver-Following Algorithm

T2 - Balancing Robustness and Efficiency in Deep Visual Detection

AU - Islam, Md Jahidul

AU - Fulton, Michael

AU - Sattar, Junaed

PY - 2019/1/1

Y1 - 2019/1/1

N2 - This letter explores the design and development of a class of robust diver detection algorithms for autonomous diver-following applications. By considering the operational challenges for underwater visual tracking in diverse real-world settings, we formulate a set of desired features of a generic diver-following algorithm. We attempt to accommodate these features and maximize general tracking performance by exploiting the state-of-the-art deep object detection models. We fine tune the building blocks of these models with a goal of balancing the tradeoff between robustness and efficiency in an on-board setting under real-time constraints. Subsequently, we design an architecturally simple convolutional neural network based diver detection model that is much faster than the state-of-the-art deep models yet provides comparable detection performances. In addition, we validate the performance and effectiveness of the proposed model through a number of diver-following experiments in closed-water and open-water environments.

AB - This letter explores the design and development of a class of robust diver detection algorithms for autonomous diver-following applications. By considering the operational challenges for underwater visual tracking in diverse real-world settings, we formulate a set of desired features of a generic diver-following algorithm. We attempt to accommodate these features and maximize general tracking performance by exploiting the state-of-the-art deep object detection models. We fine tune the building blocks of these models with a goal of balancing the tradeoff between robustness and efficiency in an on-board setting under real-time constraints. Subsequently, we design an architecturally simple convolutional neural network based diver detection model that is much faster than the state-of-the-art deep models yet provides comparable detection performances. In addition, we validate the performance and effectiveness of the proposed model through a number of diver-following experiments in closed-water and open-water environments.

KW - Human detection and tracking

KW - field robots

KW - marine robotics

UR - http://www.scopus.com/inward/record.url?scp=85063309565&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063309565&partnerID=8YFLogxK

U2 - 10.1109/LRA.2018.2882856

DO - 10.1109/LRA.2018.2882856

M3 - Article

VL - 4

SP - 113

EP - 120

JO - IEEE Robotics and Automation Letters

JF - IEEE Robotics and Automation Letters

SN - 2377-3766

IS - 1

M1 - 8543168

ER -