One Shot Similarity Metric Learning for Action Recognition

Orit Kliper-Gross; Tal Hassner; Lior Wolf

doi:https://doi.org/10.1007/978-3-642-24471-1_3

One Shot Similarity Metric Learning for Action Recognition

Orit Kliper-Gross, Tal Hassner, Lior Wolf

Weizmann Institute of Science, Open University of Israel, Tel Aviv University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The One-Shot-Similarity (OSS) is a framework for classifier-based similarity functions. It is based on the use of background samples and was shown to excel in tasks ranging from face recognition to document analysis. However, we found that its performance depends on the ability to effectively learn the underlying classifiers, which in turn depends on the underlying metric. In this work we present a metric learning technique that is geared toward improved OSS performance. We test the proposed technique using the recently presented ASLAN action similarity labeling benchmark. Enhanced, state of the art performance is obtained, and the method compares favorably to leading similarity learning techniques.

Original language	English
Title of host publication	Similarity-Based Pattern Recognition
Place of Publication	Berlin
Publisher	Springer Verlag
Pages	31-45
Number of pages	15
ISBN (Electronic)	978-3-642-24471-1
ISBN (Print)	978-3-642-24470-4, 9783642244704
DOIs	https://doi.org/10.1007/978-3-642-24471-1_3
State	Published - 2011
Event	1st International Workshop on Similarity-Based Pattern Recognition - Venice, Italy Duration: 28 Sep 2011 → 30 Sep 2011

Publication series

Name	Lecture Notes in Computer Science
Volume	7005

Conference

Conference	1st International Workshop on Similarity-Based Pattern Recognition
Abbreviated title	SIMBAD
Country/Territory	Italy
City	Venice
Period	28/09/11 → 30/09/11

Keywords

Action Similarity
Learned metrics
One-Shot-Similarity

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Access to Document

https://doi.org/10.1007/978-3-642-24471-1_3

http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.225.7028License: Unspecified

Cite this

@inproceedings{d6ac689af9fa432c89ee6ff9ae5d5110,

title = "One Shot Similarity Metric Learning for Action Recognition",

abstract = "The One-Shot-Similarity (OSS) is a framework for classifier-based similarity functions. It is based on the use of background samples and was shown to excel in tasks ranging from face recognition to document analysis. However, we found that its performance depends on the ability to effectively learn the underlying classifiers, which in turn depends on the underlying metric. In this work we present a metric learning technique that is geared toward improved OSS performance. We test the proposed technique using the recently presented ASLAN action similarity labeling benchmark. Enhanced, state of the art performance is obtained, and the method compares favorably to leading similarity learning techniques.",

keywords = "Action Similarity, Learned metrics, One-Shot-Similarity",

author = "Orit Kliper-Gross and Tal Hassner and Lior Wolf",

note = "NA; 1st International Workshop on Similarity-Based Pattern Recognition ; Conference date: 28-09-2011 Through 30-09-2011",

year = "2011",

doi = "https://doi.org/10.1007/978-3-642-24471-1_3",

language = "الإنجليزيّة",

isbn = "978-3-642-24470-4",

series = "Lecture Notes in Computer Science",

publisher = "Springer Verlag",

pages = "31--45",

booktitle = "Similarity-Based Pattern Recognition",

address = "ألمانيا",

}

TY - GEN

T1 - One Shot Similarity Metric Learning for Action Recognition

AU - Kliper-Gross, Orit

AU - Hassner, Tal

AU - Wolf, Lior

N1 - NA

PY - 2011

Y1 - 2011

N2 - The One-Shot-Similarity (OSS) is a framework for classifier-based similarity functions. It is based on the use of background samples and was shown to excel in tasks ranging from face recognition to document analysis. However, we found that its performance depends on the ability to effectively learn the underlying classifiers, which in turn depends on the underlying metric. In this work we present a metric learning technique that is geared toward improved OSS performance. We test the proposed technique using the recently presented ASLAN action similarity labeling benchmark. Enhanced, state of the art performance is obtained, and the method compares favorably to leading similarity learning techniques.

AB - The One-Shot-Similarity (OSS) is a framework for classifier-based similarity functions. It is based on the use of background samples and was shown to excel in tasks ranging from face recognition to document analysis. However, we found that its performance depends on the ability to effectively learn the underlying classifiers, which in turn depends on the underlying metric. In this work we present a metric learning technique that is geared toward improved OSS performance. We test the proposed technique using the recently presented ASLAN action similarity labeling benchmark. Enhanced, state of the art performance is obtained, and the method compares favorably to leading similarity learning techniques.

KW - Action Similarity

KW - Learned metrics

KW - One-Shot-Similarity

UR - http://www.scopus.com/inward/record.url?scp=80053374612&partnerID=8YFLogxK

U2 - https://doi.org/10.1007/978-3-642-24471-1_3

DO - https://doi.org/10.1007/978-3-642-24471-1_3

M3 - منشور من مؤتمر

SN - 978-3-642-24470-4

SN - 9783642244704

T3 - Lecture Notes in Computer Science

SP - 31

EP - 45

BT - Similarity-Based Pattern Recognition

PB - Springer Verlag

CY - Berlin

T2 - 1st International Workshop on Similarity-Based Pattern Recognition

Y2 - 28 September 2011 through 30 September 2011

ER -

One Shot Similarity Metric Learning for Action Recognition

Abstract

Publication series

Conference

Keywords

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this