Novelty Detection Using Level Set Methods with Adaptive Boundaries

Xuemei Ding, Yuhua Li, Ammar Belatreche, Liam Maguire

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

2 Citations (Scopus)

Abstract

This paper proposes a locally adaptive level setboundary description (LALSBD) method for novelty detection.The proposed method adjusts the nonlinear boundary directly inthe input space and consists of a number of processes includinglevel set function (LSF) construction, local boundary evolutionand termination. It employs kernel density estimation (KDE) to construct the LSF and form the initial boundary surrounding the training data. In order to make the boundary better fit the data distribution, a data-driven based local expanding/shrinking evolution method is proposed instead of the global evolution approach reported in our previous level set boundary description (LSBD) method. The proposed LALSBD is compared with LSBD and other four representative novelty detection methods. The experimental results demonstrate that LALSBD can detect novelevents more accurately, especially for applications which demand very high classification accuracy for normal events.
LanguageEnglish
Title of host publicationUnknown Host Publication
Pages3020-3025
Number of pages6
DOIs
Publication statusPublished - Oct 2013
Event2013 IEEE International Conference on Systems, Man, and Cybernetics - Manchester
Duration: 1 Oct 2013 → …

Conference

Conference2013 IEEE International Conference on Systems, Man, and Cybernetics
Period1/10/13 → …

Cite this

Ding, Xuemei ; Li, Yuhua ; Belatreche, Ammar ; Maguire, Liam. / Novelty Detection Using Level Set Methods with Adaptive Boundaries. Unknown Host Publication. 2013. pp. 3020-3025
@inproceedings{eac067060d2448f5958409fd29c15bd0,
title = "Novelty Detection Using Level Set Methods with Adaptive Boundaries",
abstract = "This paper proposes a locally adaptive level setboundary description (LALSBD) method for novelty detection.The proposed method adjusts the nonlinear boundary directly inthe input space and consists of a number of processes includinglevel set function (LSF) construction, local boundary evolutionand termination. It employs kernel density estimation (KDE) to construct the LSF and form the initial boundary surrounding the training data. In order to make the boundary better fit the data distribution, a data-driven based local expanding/shrinking evolution method is proposed instead of the global evolution approach reported in our previous level set boundary description (LSBD) method. The proposed LALSBD is compared with LSBD and other four representative novelty detection methods. The experimental results demonstrate that LALSBD can detect novelevents more accurately, especially for applications which demand very high classification accuracy for normal events.",
author = "Xuemei Ding and Yuhua Li and Ammar Belatreche and Liam Maguire",
year = "2013",
month = "10",
doi = "10.1109/SMC.2013.515",
language = "English",
pages = "3020--3025",
booktitle = "Unknown Host Publication",

}

Ding, X, Li, Y, Belatreche, A & Maguire, L 2013, Novelty Detection Using Level Set Methods with Adaptive Boundaries. in Unknown Host Publication. pp. 3020-3025, 2013 IEEE International Conference on Systems, Man, and Cybernetics, 1/10/13. https://doi.org/10.1109/SMC.2013.515

Novelty Detection Using Level Set Methods with Adaptive Boundaries. / Ding, Xuemei; Li, Yuhua; Belatreche, Ammar; Maguire, Liam.

Unknown Host Publication. 2013. p. 3020-3025.

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

TY - GEN

T1 - Novelty Detection Using Level Set Methods with Adaptive Boundaries

AU - Ding, Xuemei

AU - Li, Yuhua

AU - Belatreche, Ammar

AU - Maguire, Liam

PY - 2013/10

Y1 - 2013/10

N2 - This paper proposes a locally adaptive level setboundary description (LALSBD) method for novelty detection.The proposed method adjusts the nonlinear boundary directly inthe input space and consists of a number of processes includinglevel set function (LSF) construction, local boundary evolutionand termination. It employs kernel density estimation (KDE) to construct the LSF and form the initial boundary surrounding the training data. In order to make the boundary better fit the data distribution, a data-driven based local expanding/shrinking evolution method is proposed instead of the global evolution approach reported in our previous level set boundary description (LSBD) method. The proposed LALSBD is compared with LSBD and other four representative novelty detection methods. The experimental results demonstrate that LALSBD can detect novelevents more accurately, especially for applications which demand very high classification accuracy for normal events.

AB - This paper proposes a locally adaptive level setboundary description (LALSBD) method for novelty detection.The proposed method adjusts the nonlinear boundary directly inthe input space and consists of a number of processes includinglevel set function (LSF) construction, local boundary evolutionand termination. It employs kernel density estimation (KDE) to construct the LSF and form the initial boundary surrounding the training data. In order to make the boundary better fit the data distribution, a data-driven based local expanding/shrinking evolution method is proposed instead of the global evolution approach reported in our previous level set boundary description (LSBD) method. The proposed LALSBD is compared with LSBD and other four representative novelty detection methods. The experimental results demonstrate that LALSBD can detect novelevents more accurately, especially for applications which demand very high classification accuracy for normal events.

U2 - 10.1109/SMC.2013.515

DO - 10.1109/SMC.2013.515

M3 - Conference contribution

SP - 3020

EP - 3025

BT - Unknown Host Publication

ER -