Selective temporal filtering and its application to hand gesture recognition

Myung Cheol Roh, Siamac Fazli, Seong Whan Lee

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

In temporal data analysis, noisy data is inevitable in both testing and training. This noise can seriously influence the performance of the temporal data analysis. To address this problem, we propose a novel method, termed Selective Temporal Filtering that builds a noise-free model for classification during training and identifies key-feature vectors that are noise-filtered data from the input sequence during testing. The use of these key-feature vectors makes the classifier robust to noise within the input space. The proposed method is validated on a synthetic-dataset and a database of American Sign Language. Using key-feature vectors results in robust performance with respect to the noise content. Futhermore, we are able to show that the proposed method not only outperforms Conditional Random Fields and Hidden Markov Models in noisy environments, but also in a well-controlled environment where we assume no significant noise vectors exist.

Original languageEnglish
Pages (from-to)1-10
Number of pages10
JournalApplied Intelligence
DOIs
Publication statusAccepted/In press - 2016 Feb 22

Fingerprint

Gesture recognition
Testing
Hidden Markov models
Classifiers

Keywords

  • Gesture recognition
  • Key-feature vector
  • Selective Temporal Filtering

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Selective temporal filtering and its application to hand gesture recognition. / Roh, Myung Cheol; Fazli, Siamac; Lee, Seong Whan.

In: Applied Intelligence, 22.02.2016, p. 1-10.

Research output: Contribution to journalArticle

@article{1ac306970c634e2a8b9e394cb882676e,
title = "Selective temporal filtering and its application to hand gesture recognition",
abstract = "In temporal data analysis, noisy data is inevitable in both testing and training. This noise can seriously influence the performance of the temporal data analysis. To address this problem, we propose a novel method, termed Selective Temporal Filtering that builds a noise-free model for classification during training and identifies key-feature vectors that are noise-filtered data from the input sequence during testing. The use of these key-feature vectors makes the classifier robust to noise within the input space. The proposed method is validated on a synthetic-dataset and a database of American Sign Language. Using key-feature vectors results in robust performance with respect to the noise content. Futhermore, we are able to show that the proposed method not only outperforms Conditional Random Fields and Hidden Markov Models in noisy environments, but also in a well-controlled environment where we assume no significant noise vectors exist.",
keywords = "Gesture recognition, Key-feature vector, Selective Temporal Filtering",
author = "Roh, {Myung Cheol} and Siamac Fazli and Lee, {Seong Whan}",
year = "2016",
month = "2",
day = "22",
doi = "10.1007/s10489-015-0757-8",
language = "English",
pages = "1--10",
journal = "Applied Intelligence",
issn = "0924-669X",
publisher = "Springer Netherlands",

}

TY - JOUR

T1 - Selective temporal filtering and its application to hand gesture recognition

AU - Roh, Myung Cheol

AU - Fazli, Siamac

AU - Lee, Seong Whan

PY - 2016/2/22

Y1 - 2016/2/22

N2 - In temporal data analysis, noisy data is inevitable in both testing and training. This noise can seriously influence the performance of the temporal data analysis. To address this problem, we propose a novel method, termed Selective Temporal Filtering that builds a noise-free model for classification during training and identifies key-feature vectors that are noise-filtered data from the input sequence during testing. The use of these key-feature vectors makes the classifier robust to noise within the input space. The proposed method is validated on a synthetic-dataset and a database of American Sign Language. Using key-feature vectors results in robust performance with respect to the noise content. Futhermore, we are able to show that the proposed method not only outperforms Conditional Random Fields and Hidden Markov Models in noisy environments, but also in a well-controlled environment where we assume no significant noise vectors exist.

AB - In temporal data analysis, noisy data is inevitable in both testing and training. This noise can seriously influence the performance of the temporal data analysis. To address this problem, we propose a novel method, termed Selective Temporal Filtering that builds a noise-free model for classification during training and identifies key-feature vectors that are noise-filtered data from the input sequence during testing. The use of these key-feature vectors makes the classifier robust to noise within the input space. The proposed method is validated on a synthetic-dataset and a database of American Sign Language. Using key-feature vectors results in robust performance with respect to the noise content. Futhermore, we are able to show that the proposed method not only outperforms Conditional Random Fields and Hidden Markov Models in noisy environments, but also in a well-controlled environment where we assume no significant noise vectors exist.

KW - Gesture recognition

KW - Key-feature vector

KW - Selective Temporal Filtering

UR - http://www.scopus.com/inward/record.url?scp=84959149303&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84959149303&partnerID=8YFLogxK

U2 - 10.1007/s10489-015-0757-8

DO - 10.1007/s10489-015-0757-8

M3 - Article

AN - SCOPUS:84959149303

SP - 1

EP - 10

JO - Applied Intelligence

JF - Applied Intelligence

SN - 0924-669X

ER -