CERIAS - Center for Education and Research in Information Assurance and Security

Skip Navigation
CERIAS Logo
Purdue University
Center for Education and Research in Information Assurance and Security

TC: Medium: Security and Privacy Preserving Data Mining and Management for Distributed Domains

Principal Investigator: Arif Ghafoor

A fundamental but challenging issue in information security is secure sharing and management of sensitive data and information among numerous organizations that form large-scale e-enterprises. Today, an increasing number of enterprises are using the Internet for managing and sharing users’ and enterprise information through online databases. However, security and privacy of data is an overriding concern currently limiting the proliferation of information technology. Among others, the two key security issues are; (a) interoperability challenge of diverse security policies exercised by collaborating organizations for sharing sensitive information and; (b) using collaborative knowledge for detecting and responding to any emerging security or emergency threats. The objective of this project to develop a scalable multi-domain information security framework which: (a) facilitates large-scale integration, mining, and analysis of multimedia data residing in multiple domains for identifying sensitive information that require controlled dissemination (b) allows distributed real-time content analysis of multimedia data streams across multiple-domains for correlating detectable events, (c)l provides integration and evolution of inter-domain access control policies for distributed multimedia applications, and (d) ensures protection of user’s information.

The project uses several innovative approaches to meet theses objectives. First, novel real-time data mining and stream-correlation methods is used for extracting useful content and detecting events embedded in data streams by using multiple sliding window operators. Second, an efficient privacy/security based multimedia data clustering mechanism allows a prevailing security policy to be adaptable to the changing context and the datasets. The uniqueness of this approach is the adaptive nature of security policies. In addition, a novel XML-based integrated mechanism is used for encoding of multimedia data in conjunction with multi-domain security policies for ensuring technical viability and portability of the system.

The proposed research is expected to have direct and long-term impact on developing secure information infrastructures, and in general, meeting the national information infrastructure security needs and threat management.

Personnel

Other PIs: Walid Aref Ahmed Elmagarmid

Students: Zahid Pervaiz Ihsanulhaq Sarfraz Amana Arshad

Keywords: content analysis., data mining, information security, multimedia databases, Privacy