Deep correlated metric learning for sketch-based 3D shape retrieval

Guoxian Dai, Jin Xie, Fan Zhu, Yi Fang

Research output: Contribution to conferencePaper

Abstract

The explosive growth of 3D models has led to the pressing demand for an efficient searching system. Traditional modelbased search is usually not convenient, since people don't always have 3D model available by side. The sketch-based 3D shape retrieval is a promising candidate due to its simple-ness and efficiency. The main challenge for sketch-based 3D shape retrieval is the discrepancy across different domains. In the paper, we propose a novel deep correlated metric learning (DCML) method to mitigate the discrepancy between sketch and 3D shape domains. The proposed DCML trains two distinct deep neural networks (one for each domain) jointly with one loss, which learns two deep nonlinear transformations to map features from both domains into a nonlinear feature space. The proposed loss, including discriminative loss and correlation loss, aims to increase the discrimination of features within each domain as well as the correlation between different domains. In the transfered space, the discriminative loss minimizes the intra-class distance of the deep transformed features and maximizes the inter-class distance of the deep transformed features at least a predefined margin within each domain, while the correlation loss focuses on minimizing the distribution discrepancy across different domains. Our proposed method is evaluated on SHREC 2013 and 2014 benchmarks, and the experimental results demonstrate the superiority of our proposed method over the state-of-the-art methods.

Original languageEnglish (US)
Pages4002-4008
Number of pages7
StatePublished - Jan 1 2017
Event31st AAAI Conference on Artificial Intelligence, AAAI 2017 - San Francisco, United States
Duration: Feb 4 2017Feb 10 2017

Other

Other31st AAAI Conference on Artificial Intelligence, AAAI 2017
CountryUnited States
CitySan Francisco
Period2/4/172/10/17

Fingerprint

Deep neural networks

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Dai, G., Xie, J., Zhu, F., & Fang, Y. (2017). Deep correlated metric learning for sketch-based 3D shape retrieval. 4002-4008. Paper presented at 31st AAAI Conference on Artificial Intelligence, AAAI 2017, San Francisco, United States.

Deep correlated metric learning for sketch-based 3D shape retrieval. / Dai, Guoxian; Xie, Jin; Zhu, Fan; Fang, Yi.

2017. 4002-4008 Paper presented at 31st AAAI Conference on Artificial Intelligence, AAAI 2017, San Francisco, United States.

Research output: Contribution to conferencePaper

Dai, G, Xie, J, Zhu, F & Fang, Y 2017, 'Deep correlated metric learning for sketch-based 3D shape retrieval' Paper presented at 31st AAAI Conference on Artificial Intelligence, AAAI 2017, San Francisco, United States, 2/4/17 - 2/10/17, pp. 4002-4008.
Dai G, Xie J, Zhu F, Fang Y. Deep correlated metric learning for sketch-based 3D shape retrieval. 2017. Paper presented at 31st AAAI Conference on Artificial Intelligence, AAAI 2017, San Francisco, United States.
Dai, Guoxian ; Xie, Jin ; Zhu, Fan ; Fang, Yi. / Deep correlated metric learning for sketch-based 3D shape retrieval. Paper presented at 31st AAAI Conference on Artificial Intelligence, AAAI 2017, San Francisco, United States.7 p.
@conference{99e36420dddd4a6e8cf88b9a20612658,
title = "Deep correlated metric learning for sketch-based 3D shape retrieval",
abstract = "The explosive growth of 3D models has led to the pressing demand for an efficient searching system. Traditional modelbased search is usually not convenient, since people don't always have 3D model available by side. The sketch-based 3D shape retrieval is a promising candidate due to its simple-ness and efficiency. The main challenge for sketch-based 3D shape retrieval is the discrepancy across different domains. In the paper, we propose a novel deep correlated metric learning (DCML) method to mitigate the discrepancy between sketch and 3D shape domains. The proposed DCML trains two distinct deep neural networks (one for each domain) jointly with one loss, which learns two deep nonlinear transformations to map features from both domains into a nonlinear feature space. The proposed loss, including discriminative loss and correlation loss, aims to increase the discrimination of features within each domain as well as the correlation between different domains. In the transfered space, the discriminative loss minimizes the intra-class distance of the deep transformed features and maximizes the inter-class distance of the deep transformed features at least a predefined margin within each domain, while the correlation loss focuses on minimizing the distribution discrepancy across different domains. Our proposed method is evaluated on SHREC 2013 and 2014 benchmarks, and the experimental results demonstrate the superiority of our proposed method over the state-of-the-art methods.",
author = "Guoxian Dai and Jin Xie and Fan Zhu and Yi Fang",
year = "2017",
month = "1",
day = "1",
language = "English (US)",
pages = "4002--4008",
note = "31st AAAI Conference on Artificial Intelligence, AAAI 2017 ; Conference date: 04-02-2017 Through 10-02-2017",

}

TY - CONF

T1 - Deep correlated metric learning for sketch-based 3D shape retrieval

AU - Dai, Guoxian

AU - Xie, Jin

AU - Zhu, Fan

AU - Fang, Yi

PY - 2017/1/1

Y1 - 2017/1/1

N2 - The explosive growth of 3D models has led to the pressing demand for an efficient searching system. Traditional modelbased search is usually not convenient, since people don't always have 3D model available by side. The sketch-based 3D shape retrieval is a promising candidate due to its simple-ness and efficiency. The main challenge for sketch-based 3D shape retrieval is the discrepancy across different domains. In the paper, we propose a novel deep correlated metric learning (DCML) method to mitigate the discrepancy between sketch and 3D shape domains. The proposed DCML trains two distinct deep neural networks (one for each domain) jointly with one loss, which learns two deep nonlinear transformations to map features from both domains into a nonlinear feature space. The proposed loss, including discriminative loss and correlation loss, aims to increase the discrimination of features within each domain as well as the correlation between different domains. In the transfered space, the discriminative loss minimizes the intra-class distance of the deep transformed features and maximizes the inter-class distance of the deep transformed features at least a predefined margin within each domain, while the correlation loss focuses on minimizing the distribution discrepancy across different domains. Our proposed method is evaluated on SHREC 2013 and 2014 benchmarks, and the experimental results demonstrate the superiority of our proposed method over the state-of-the-art methods.

AB - The explosive growth of 3D models has led to the pressing demand for an efficient searching system. Traditional modelbased search is usually not convenient, since people don't always have 3D model available by side. The sketch-based 3D shape retrieval is a promising candidate due to its simple-ness and efficiency. The main challenge for sketch-based 3D shape retrieval is the discrepancy across different domains. In the paper, we propose a novel deep correlated metric learning (DCML) method to mitigate the discrepancy between sketch and 3D shape domains. The proposed DCML trains two distinct deep neural networks (one for each domain) jointly with one loss, which learns two deep nonlinear transformations to map features from both domains into a nonlinear feature space. The proposed loss, including discriminative loss and correlation loss, aims to increase the discrimination of features within each domain as well as the correlation between different domains. In the transfered space, the discriminative loss minimizes the intra-class distance of the deep transformed features and maximizes the inter-class distance of the deep transformed features at least a predefined margin within each domain, while the correlation loss focuses on minimizing the distribution discrepancy across different domains. Our proposed method is evaluated on SHREC 2013 and 2014 benchmarks, and the experimental results demonstrate the superiority of our proposed method over the state-of-the-art methods.

UR - http://www.scopus.com/inward/record.url?scp=85030481364&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85030481364&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:85030481364

SP - 4002

EP - 4008

ER -