Abstract

In the recent years, deep learning algorithms have achieved state-of-the-art performance in a variety of computer vision applications. In this paper, we propose a novel active learning framework to select the most informative unlabeled samples to train a deep belief network model. We introduce a loss function specific to the active learning task and train the model to minimize the loss function. To the best of our knowledge, this is the first research effort to integrate an active learning based criterion in the loss function used to train a deep belief network. Our extensive empirical studies on a wide variety of uni-modal and multi-modal vision datasets corroborate the potential of the method for real-world image recognition applications.

Original languageEnglish (US)
Title of host publication2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings
PublisherIEEE Computer Society
Pages3934-3938
Number of pages5
Volume2017-September
ISBN (Electronic)9781509021758
DOIs
StatePublished - Feb 20 2018
Event24th IEEE International Conference on Image Processing, ICIP 2017 - Beijing, China
Duration: Sep 17 2017Sep 20 2017

Other

Other24th IEEE International Conference on Image Processing, ICIP 2017
CountryChina
CityBeijing
Period9/17/179/20/17

Fingerprint

Image classification
Bayesian networks
Image recognition
Learning algorithms
Computer vision
Problem-Based Learning

Keywords

  • Active learning
  • Computer vision
  • Deep belief networks
  • Deep learning
  • Entropy

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Ranganathan, H., Demakethepalli Venkateswara, H., Chakraborty, S., & Panchanathan, S. (2018). Deep active learning for image classification. In 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings (Vol. 2017-September, pp. 3934-3938). IEEE Computer Society. https://doi.org/10.1109/ICIP.2017.8297020

Deep active learning for image classification. / Ranganathan, Hiranmayi; Demakethepalli Venkateswara, Hemanth; Chakraborty, Shayok; Panchanathan, Sethuraman.

2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. Vol. 2017-September IEEE Computer Society, 2018. p. 3934-3938.

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

Ranganathan, H, Demakethepalli Venkateswara, H, Chakraborty, S & Panchanathan, S 2018, Deep active learning for image classification. in 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. vol. 2017-September, IEEE Computer Society, pp. 3934-3938, 24th IEEE International Conference on Image Processing, ICIP 2017, Beijing, China, 9/17/17. https://doi.org/10.1109/ICIP.2017.8297020
Ranganathan H, Demakethepalli Venkateswara H, Chakraborty S, Panchanathan S. Deep active learning for image classification. In 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. Vol. 2017-September. IEEE Computer Society. 2018. p. 3934-3938 https://doi.org/10.1109/ICIP.2017.8297020
Ranganathan, Hiranmayi ; Demakethepalli Venkateswara, Hemanth ; Chakraborty, Shayok ; Panchanathan, Sethuraman. / Deep active learning for image classification. 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. Vol. 2017-September IEEE Computer Society, 2018. pp. 3934-3938
@inproceedings{deab8110ab414c09a5880882adb0b503,
title = "Deep active learning for image classification",
abstract = "In the recent years, deep learning algorithms have achieved state-of-the-art performance in a variety of computer vision applications. In this paper, we propose a novel active learning framework to select the most informative unlabeled samples to train a deep belief network model. We introduce a loss function specific to the active learning task and train the model to minimize the loss function. To the best of our knowledge, this is the first research effort to integrate an active learning based criterion in the loss function used to train a deep belief network. Our extensive empirical studies on a wide variety of uni-modal and multi-modal vision datasets corroborate the potential of the method for real-world image recognition applications.",
keywords = "Active learning, Computer vision, Deep belief networks, Deep learning, Entropy",
author = "Hiranmayi Ranganathan and {Demakethepalli Venkateswara}, Hemanth and Shayok Chakraborty and Sethuraman Panchanathan",
year = "2018",
month = "2",
day = "20",
doi = "10.1109/ICIP.2017.8297020",
language = "English (US)",
volume = "2017-September",
pages = "3934--3938",
booktitle = "2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Deep active learning for image classification

AU - Ranganathan, Hiranmayi

AU - Demakethepalli Venkateswara, Hemanth

AU - Chakraborty, Shayok

AU - Panchanathan, Sethuraman

PY - 2018/2/20

Y1 - 2018/2/20

N2 - In the recent years, deep learning algorithms have achieved state-of-the-art performance in a variety of computer vision applications. In this paper, we propose a novel active learning framework to select the most informative unlabeled samples to train a deep belief network model. We introduce a loss function specific to the active learning task and train the model to minimize the loss function. To the best of our knowledge, this is the first research effort to integrate an active learning based criterion in the loss function used to train a deep belief network. Our extensive empirical studies on a wide variety of uni-modal and multi-modal vision datasets corroborate the potential of the method for real-world image recognition applications.

AB - In the recent years, deep learning algorithms have achieved state-of-the-art performance in a variety of computer vision applications. In this paper, we propose a novel active learning framework to select the most informative unlabeled samples to train a deep belief network model. We introduce a loss function specific to the active learning task and train the model to minimize the loss function. To the best of our knowledge, this is the first research effort to integrate an active learning based criterion in the loss function used to train a deep belief network. Our extensive empirical studies on a wide variety of uni-modal and multi-modal vision datasets corroborate the potential of the method for real-world image recognition applications.

KW - Active learning

KW - Computer vision

KW - Deep belief networks

KW - Deep learning

KW - Entropy

UR - http://www.scopus.com/inward/record.url?scp=85045307496&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045307496&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2017.8297020

DO - 10.1109/ICIP.2017.8297020

M3 - Conference contribution

AN - SCOPUS:85045307496

VL - 2017-September

SP - 3934

EP - 3938

BT - 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings

PB - IEEE Computer Society

ER -