Video saliency detection based on spatiotemporal feature learning

Se Ho Lee, Jin Hwan Kim, Kwang Pyo Choi, Jae Young Sim, Chang-Su Kim

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

15 Citations (Scopus)

Abstract

A video saliency detection algorithm based on feature learning, called ROCT, is proposed in this work. To detect salient regions, we design multiple spatiotemporal features and combine those features using a support vector machine (SVM). We extract the spatial features of rarity, compactness, and center prior by analyzing the color distribution in each image frame. Also, we obtain the temporal features of motion intensity and motion contrast to identify visually important motions. We train an SVM classifier using the spatiotemporal features extracted from training video sequences. Finally, we compute the visual saliency of each patch in an input sequence using the trained classifier. Experimental results demonstrate that the proposed algorithm provides more accurate and reliable results of saliency detection than conventional algorithms.

Original languageEnglish
Title of host publication2014 IEEE International Conference on Image Processing, ICIP 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1120-1124
Number of pages5
ISBN (Print)9781479957514
DOIs
Publication statusPublished - 2014 Jan 28

Fingerprint

Support vector machines
Classifiers
Color

Keywords

  • machine learning
  • spatiotemporal features
  • support vector machine
  • Video saliency detection

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Lee, S. H., Kim, J. H., Choi, K. P., Sim, J. Y., & Kim, C-S. (2014). Video saliency detection based on spatiotemporal feature learning. In 2014 IEEE International Conference on Image Processing, ICIP 2014 (pp. 1120-1124). [7025223] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICIP.2014.7025223

Video saliency detection based on spatiotemporal feature learning. / Lee, Se Ho; Kim, Jin Hwan; Choi, Kwang Pyo; Sim, Jae Young; Kim, Chang-Su.

2014 IEEE International Conference on Image Processing, ICIP 2014. Institute of Electrical and Electronics Engineers Inc., 2014. p. 1120-1124 7025223.

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

Lee, SH, Kim, JH, Choi, KP, Sim, JY & Kim, C-S 2014, Video saliency detection based on spatiotemporal feature learning. in 2014 IEEE International Conference on Image Processing, ICIP 2014., 7025223, Institute of Electrical and Electronics Engineers Inc., pp. 1120-1124. https://doi.org/10.1109/ICIP.2014.7025223
Lee SH, Kim JH, Choi KP, Sim JY, Kim C-S. Video saliency detection based on spatiotemporal feature learning. In 2014 IEEE International Conference on Image Processing, ICIP 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 1120-1124. 7025223 https://doi.org/10.1109/ICIP.2014.7025223
Lee, Se Ho ; Kim, Jin Hwan ; Choi, Kwang Pyo ; Sim, Jae Young ; Kim, Chang-Su. / Video saliency detection based on spatiotemporal feature learning. 2014 IEEE International Conference on Image Processing, ICIP 2014. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 1120-1124
@inproceedings{f81e938d800f493e8e0dd68ad211602d,
title = "Video saliency detection based on spatiotemporal feature learning",
abstract = "A video saliency detection algorithm based on feature learning, called ROCT, is proposed in this work. To detect salient regions, we design multiple spatiotemporal features and combine those features using a support vector machine (SVM). We extract the spatial features of rarity, compactness, and center prior by analyzing the color distribution in each image frame. Also, we obtain the temporal features of motion intensity and motion contrast to identify visually important motions. We train an SVM classifier using the spatiotemporal features extracted from training video sequences. Finally, we compute the visual saliency of each patch in an input sequence using the trained classifier. Experimental results demonstrate that the proposed algorithm provides more accurate and reliable results of saliency detection than conventional algorithms.",
keywords = "machine learning, spatiotemporal features, support vector machine, Video saliency detection",
author = "Lee, {Se Ho} and Kim, {Jin Hwan} and Choi, {Kwang Pyo} and Sim, {Jae Young} and Chang-Su Kim",
year = "2014",
month = "1",
day = "28",
doi = "10.1109/ICIP.2014.7025223",
language = "English",
isbn = "9781479957514",
pages = "1120--1124",
booktitle = "2014 IEEE International Conference on Image Processing, ICIP 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Video saliency detection based on spatiotemporal feature learning

AU - Lee, Se Ho

AU - Kim, Jin Hwan

AU - Choi, Kwang Pyo

AU - Sim, Jae Young

AU - Kim, Chang-Su

PY - 2014/1/28

Y1 - 2014/1/28

N2 - A video saliency detection algorithm based on feature learning, called ROCT, is proposed in this work. To detect salient regions, we design multiple spatiotemporal features and combine those features using a support vector machine (SVM). We extract the spatial features of rarity, compactness, and center prior by analyzing the color distribution in each image frame. Also, we obtain the temporal features of motion intensity and motion contrast to identify visually important motions. We train an SVM classifier using the spatiotemporal features extracted from training video sequences. Finally, we compute the visual saliency of each patch in an input sequence using the trained classifier. Experimental results demonstrate that the proposed algorithm provides more accurate and reliable results of saliency detection than conventional algorithms.

AB - A video saliency detection algorithm based on feature learning, called ROCT, is proposed in this work. To detect salient regions, we design multiple spatiotemporal features and combine those features using a support vector machine (SVM). We extract the spatial features of rarity, compactness, and center prior by analyzing the color distribution in each image frame. Also, we obtain the temporal features of motion intensity and motion contrast to identify visually important motions. We train an SVM classifier using the spatiotemporal features extracted from training video sequences. Finally, we compute the visual saliency of each patch in an input sequence using the trained classifier. Experimental results demonstrate that the proposed algorithm provides more accurate and reliable results of saliency detection than conventional algorithms.

KW - machine learning

KW - spatiotemporal features

KW - support vector machine

KW - Video saliency detection

UR - http://www.scopus.com/inward/record.url?scp=84943759090&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84943759090&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2014.7025223

DO - 10.1109/ICIP.2014.7025223

M3 - Conference contribution

SN - 9781479957514

SP - 1120

EP - 1124

BT - 2014 IEEE International Conference on Image Processing, ICIP 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -