Multi-scale visual attention & saliency modelling with decision theory

A.C. Le Ngo, L.-M. Ang, G. Qiu, K.P. Seng

Research output: Book chapter/Published conference paperConference paper

Abstract

Recently, an information-based saliency technique which is biologically plausible and computationally feasible called Discriminant Saliency (DIS) has been proposed. While DIS successfully defines discriminant saliency in the information theoretic sense, its implementation restraints the sampled features to a single fixed-size window and creates a bias towards objects with distinctive features fitted in the window size. This paper proposes a multi-scale discriminant saliency (MDIS) technique for visual attention which uses the wavelet transform for the multi-resolution framework. MDIS utilizes mutual information between classes and feature distribution to quantify classifying discriminant power as saliency value in multiple dyadic-scale structures. The paper will present simulations on Neil Bruce's image database with quantitative and qualitative results showing the advantages of MDIS over DIS. For quantitative comparisons, numerical tests AUC, NSS, LCC are measured and several plots are generated to visualized differences between simulation modes; meanwhile, qualitative evaluation is a visual examination of synthesized saliency maps of general natural scenes. © 2013 IEEE.
Original languageUndefined/Unknown
Title of host publication2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings
Place of PublicationUnited States
PublisherIEEE, Institute of Electrical and Electronics Engineers
Pages216-220
ISBN (Electronic)9781479923410
DOIs
Publication statusPublished - 13 Feb 2014
Event2013 20th IEEE International Conference on Image Processing: ICIP 2013 - Melbourne Convention and Exhibition Centre, Melbourne, Australia
Duration: 15 Sep 201318 Sep 2013
https://www2.securecms.com/ICIP2013/

Conference

Conference2013 20th IEEE International Conference on Image Processing
CountryAustralia
CityMelbourne
Period15/09/1318/09/13
OtherThe International Conference on Image Processing (ICIP), sponsored by the IEEE Signal Processing Society, is the premier forum for the presentation of technological advances and research results in the fields of theoretical, experimental, and applied image and video processing. ICIP 2013, the twentieth in the series that has been held annually since 1994, brings together leading engineers and scientists in image and video processing from around the world.
Internet address

Cite this

Le Ngo, A. C., Ang, L-M., Qiu, G., & Seng, K. P. (2014). Multi-scale visual attention & saliency modelling with decision theory. In 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings (pp. 216-220). United States: IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICIP.2013.6738045
Le Ngo, A.C. ; Ang, L.-M. ; Qiu, G. ; Seng, K.P. / Multi-scale visual attention & saliency modelling with decision theory. 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings. United States : IEEE, Institute of Electrical and Electronics Engineers, 2014. pp. 216-220
@inproceedings{ead174f7afa54e4bb36693453ff665c9,
title = "Multi-scale visual attention & saliency modelling with decision theory",
abstract = "Recently, an information-based saliency technique which is biologically plausible and computationally feasible called Discriminant Saliency (DIS) has been proposed. While DIS successfully defines discriminant saliency in the information theoretic sense, its implementation restraints the sampled features to a single fixed-size window and creates a bias towards objects with distinctive features fitted in the window size. This paper proposes a multi-scale discriminant saliency (MDIS) technique for visual attention which uses the wavelet transform for the multi-resolution framework. MDIS utilizes mutual information between classes and feature distribution to quantify classifying discriminant power as saliency value in multiple dyadic-scale structures. The paper will present simulations on Neil Bruce's image database with quantitative and qualitative results showing the advantages of MDIS over DIS. For quantitative comparisons, numerical tests AUC, NSS, LCC are measured and several plots are generated to visualized differences between simulation modes; meanwhile, qualitative evaluation is a visual examination of synthesized saliency maps of general natural scenes. {\circledC} 2013 IEEE.",
keywords = "Visualization, Hidden Markov models, Equations, Educational institutions, Estimation, Image segmentation, Mathematical model",
author = "{Le Ngo}, A.C. and L.-M. Ang and G. Qiu and K.P. Seng",
note = "cited By 0",
year = "2014",
month = "2",
day = "13",
doi = "10.1109/ICIP.2013.6738045",
language = "Undefined/Unknown",
pages = "216--220",
booktitle = "2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
address = "United States",

}

Le Ngo, AC, Ang, L-M, Qiu, G & Seng, KP 2014, Multi-scale visual attention & saliency modelling with decision theory. in 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings. IEEE, Institute of Electrical and Electronics Engineers, United States, pp. 216-220, 2013 20th IEEE International Conference on Image Processing, Melbourne, Australia, 15/09/13. https://doi.org/10.1109/ICIP.2013.6738045

Multi-scale visual attention & saliency modelling with decision theory. / Le Ngo, A.C.; Ang, L.-M.; Qiu, G.; Seng, K.P.

2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings. United States : IEEE, Institute of Electrical and Electronics Engineers, 2014. p. 216-220.

Research output: Book chapter/Published conference paperConference paper

TY - GEN

T1 - Multi-scale visual attention & saliency modelling with decision theory

AU - Le Ngo, A.C.

AU - Ang, L.-M.

AU - Qiu, G.

AU - Seng, K.P.

N1 - cited By 0

PY - 2014/2/13

Y1 - 2014/2/13

N2 - Recently, an information-based saliency technique which is biologically plausible and computationally feasible called Discriminant Saliency (DIS) has been proposed. While DIS successfully defines discriminant saliency in the information theoretic sense, its implementation restraints the sampled features to a single fixed-size window and creates a bias towards objects with distinctive features fitted in the window size. This paper proposes a multi-scale discriminant saliency (MDIS) technique for visual attention which uses the wavelet transform for the multi-resolution framework. MDIS utilizes mutual information between classes and feature distribution to quantify classifying discriminant power as saliency value in multiple dyadic-scale structures. The paper will present simulations on Neil Bruce's image database with quantitative and qualitative results showing the advantages of MDIS over DIS. For quantitative comparisons, numerical tests AUC, NSS, LCC are measured and several plots are generated to visualized differences between simulation modes; meanwhile, qualitative evaluation is a visual examination of synthesized saliency maps of general natural scenes. © 2013 IEEE.

AB - Recently, an information-based saliency technique which is biologically plausible and computationally feasible called Discriminant Saliency (DIS) has been proposed. While DIS successfully defines discriminant saliency in the information theoretic sense, its implementation restraints the sampled features to a single fixed-size window and creates a bias towards objects with distinctive features fitted in the window size. This paper proposes a multi-scale discriminant saliency (MDIS) technique for visual attention which uses the wavelet transform for the multi-resolution framework. MDIS utilizes mutual information between classes and feature distribution to quantify classifying discriminant power as saliency value in multiple dyadic-scale structures. The paper will present simulations on Neil Bruce's image database with quantitative and qualitative results showing the advantages of MDIS over DIS. For quantitative comparisons, numerical tests AUC, NSS, LCC are measured and several plots are generated to visualized differences between simulation modes; meanwhile, qualitative evaluation is a visual examination of synthesized saliency maps of general natural scenes. © 2013 IEEE.

KW - Visualization

KW - Hidden Markov models

KW - Equations

KW - Educational institutions

KW - Estimation

KW - Image segmentation

KW - Mathematical model

UR - http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6737998

U2 - 10.1109/ICIP.2013.6738045

DO - 10.1109/ICIP.2013.6738045

M3 - Conference paper

SP - 216

EP - 220

BT - 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - United States

ER -

Le Ngo AC, Ang L-M, Qiu G, Seng KP. Multi-scale visual attention & saliency modelling with decision theory. In 2013 IEEE International Conference on Image Processing, ICIP 2013 - Proceedings. United States: IEEE, Institute of Electrical and Electronics Engineers. 2014. p. 216-220 https://doi.org/10.1109/ICIP.2013.6738045