DPcode: Privacy-Preserving Frequent Visual Patterns Publication on Cloud

Zhan Qin, Kui Ren, Ting Yu, Jian Weng

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Nowadays, cloud has become a promising multimedia data processing and sharing platform. Many institutes and companies plan to outsource and share their large-scale video and image datasets on cloud for scientific research and public interest. Among various video applications, the discovery of frequent visual patterns over graphical data is an exploratory and important technique. However, the privacy concerns over the leakage of sensitive information contained in the videos/images impedes the further implementation. Although the frequent visual patterns mining (FVPM) algorithm aggregates summary over individual frames and seems not to pose privacy threat, the private information contained in individual frames still may be leaked from the statistical result. In this paper, we study the problem of privacy-preserving publishing of graphical data FVPM on cloud. We propose the first differentially private frequent visual patterns mining algorithm for graphical data, named DPcode. We propose a novel mechanism that integrates the privacy-preserving visual word conversion with the differentially private mechanism under the noise allocation strategy of the sparse vector technique. The optimized algorithms properly allocate the privacy budgets among different phases in FPM algorithm over images and reduce the corresponding data distortion. Extensive experiments are conducted based on datasets commonly used in visual mining algorithms. The results show that our approach achieves high utility while satisfying a practical privacy requirement.

Original languageEnglish
Article number7422120
Pages (from-to)929-939
Number of pages11
JournalIEEE Transactions on Multimedia
Volume18
Issue number5
DOIs
Publication statusPublished - 1 May 2016

Fingerprint

Industry
Experiments

Keywords

  • cloud
  • data publication
  • differential privacy
  • Frequent pattern mining

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Signal Processing
  • Media Technology
  • Computer Science Applications

Cite this

DPcode : Privacy-Preserving Frequent Visual Patterns Publication on Cloud. / Qin, Zhan; Ren, Kui; Yu, Ting; Weng, Jian.

In: IEEE Transactions on Multimedia, Vol. 18, No. 5, 7422120, 01.05.2016, p. 929-939.

Research output: Contribution to journalArticle

Qin, Zhan ; Ren, Kui ; Yu, Ting ; Weng, Jian. / DPcode : Privacy-Preserving Frequent Visual Patterns Publication on Cloud. In: IEEE Transactions on Multimedia. 2016 ; Vol. 18, No. 5. pp. 929-939.
@article{069dc2d92c5d43a7993ed2f0124a2236,
title = "DPcode: Privacy-Preserving Frequent Visual Patterns Publication on Cloud",
abstract = "Nowadays, cloud has become a promising multimedia data processing and sharing platform. Many institutes and companies plan to outsource and share their large-scale video and image datasets on cloud for scientific research and public interest. Among various video applications, the discovery of frequent visual patterns over graphical data is an exploratory and important technique. However, the privacy concerns over the leakage of sensitive information contained in the videos/images impedes the further implementation. Although the frequent visual patterns mining (FVPM) algorithm aggregates summary over individual frames and seems not to pose privacy threat, the private information contained in individual frames still may be leaked from the statistical result. In this paper, we study the problem of privacy-preserving publishing of graphical data FVPM on cloud. We propose the first differentially private frequent visual patterns mining algorithm for graphical data, named DPcode. We propose a novel mechanism that integrates the privacy-preserving visual word conversion with the differentially private mechanism under the noise allocation strategy of the sparse vector technique. The optimized algorithms properly allocate the privacy budgets among different phases in FPM algorithm over images and reduce the corresponding data distortion. Extensive experiments are conducted based on datasets commonly used in visual mining algorithms. The results show that our approach achieves high utility while satisfying a practical privacy requirement.",
keywords = "cloud, data publication, differential privacy, Frequent pattern mining",
author = "Zhan Qin and Kui Ren and Ting Yu and Jian Weng",
year = "2016",
month = "5",
day = "1",
doi = "10.1109/TMM.2016.2535729",
language = "English",
volume = "18",
pages = "929--939",
journal = "IEEE Transactions on Multimedia",
issn = "1520-9210",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - DPcode

T2 - Privacy-Preserving Frequent Visual Patterns Publication on Cloud

AU - Qin, Zhan

AU - Ren, Kui

AU - Yu, Ting

AU - Weng, Jian

PY - 2016/5/1

Y1 - 2016/5/1

N2 - Nowadays, cloud has become a promising multimedia data processing and sharing platform. Many institutes and companies plan to outsource and share their large-scale video and image datasets on cloud for scientific research and public interest. Among various video applications, the discovery of frequent visual patterns over graphical data is an exploratory and important technique. However, the privacy concerns over the leakage of sensitive information contained in the videos/images impedes the further implementation. Although the frequent visual patterns mining (FVPM) algorithm aggregates summary over individual frames and seems not to pose privacy threat, the private information contained in individual frames still may be leaked from the statistical result. In this paper, we study the problem of privacy-preserving publishing of graphical data FVPM on cloud. We propose the first differentially private frequent visual patterns mining algorithm for graphical data, named DPcode. We propose a novel mechanism that integrates the privacy-preserving visual word conversion with the differentially private mechanism under the noise allocation strategy of the sparse vector technique. The optimized algorithms properly allocate the privacy budgets among different phases in FPM algorithm over images and reduce the corresponding data distortion. Extensive experiments are conducted based on datasets commonly used in visual mining algorithms. The results show that our approach achieves high utility while satisfying a practical privacy requirement.

AB - Nowadays, cloud has become a promising multimedia data processing and sharing platform. Many institutes and companies plan to outsource and share their large-scale video and image datasets on cloud for scientific research and public interest. Among various video applications, the discovery of frequent visual patterns over graphical data is an exploratory and important technique. However, the privacy concerns over the leakage of sensitive information contained in the videos/images impedes the further implementation. Although the frequent visual patterns mining (FVPM) algorithm aggregates summary over individual frames and seems not to pose privacy threat, the private information contained in individual frames still may be leaked from the statistical result. In this paper, we study the problem of privacy-preserving publishing of graphical data FVPM on cloud. We propose the first differentially private frequent visual patterns mining algorithm for graphical data, named DPcode. We propose a novel mechanism that integrates the privacy-preserving visual word conversion with the differentially private mechanism under the noise allocation strategy of the sparse vector technique. The optimized algorithms properly allocate the privacy budgets among different phases in FPM algorithm over images and reduce the corresponding data distortion. Extensive experiments are conducted based on datasets commonly used in visual mining algorithms. The results show that our approach achieves high utility while satisfying a practical privacy requirement.

KW - cloud

KW - data publication

KW - differential privacy

KW - Frequent pattern mining

UR - http://www.scopus.com/inward/record.url?scp=84968799905&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84968799905&partnerID=8YFLogxK

U2 - 10.1109/TMM.2016.2535729

DO - 10.1109/TMM.2016.2535729

M3 - Article

AN - SCOPUS:84968799905

VL - 18

SP - 929

EP - 939

JO - IEEE Transactions on Multimedia

JF - IEEE Transactions on Multimedia

SN - 1520-9210

IS - 5

M1 - 7422120

ER -