When large images are used for big data analysis, they impose new challenges in protecting image privacy. For example, a geographic image may consist of several sensitive areas or layers. When it is uploaded into servers, the image will be accessed by diverse subjects. Traditional access-control methods regulate access privileges to a single image, and their access-control strategies are stored in servers, which impose two shortcomings: (1) fine-grained access control is not guaranteed for areas/layers in a single image that needs to maintain secret for different roles and (2) accesscontrol policies that are stored in servers suffer from multiple attacks (e.g., transferring attacks). In this chapter, we propose a novel watermark-based access-control model in which access-control policies are associated with objects being accessed (called an in situ model). The proposed model integrates access-control policies as watermarks within images, without relying on the availability of servers or connecting networks. The access control for images is still maintained even though images are redistributed again to further subjects. Therefore, access-control policies can be delivered together with the big data of images. Moreover, we propose a hierarchical key-role-area model for fine-grained encryption, especially for large-sized images such as geographic maps. The extensive analysis justifies the security and performance of the proposed model.

Chapter Contents:

• 4.1 Preliminaries
• 4.2 Introduction
• 4.3 Related work
• 4.4 Problem formulation
• 4.4.1 System model
• 4.4.2 Attack models
• 4.4.2.1 Transferring attack
• 4.4.2.2 Distributed denial of service (DDoS) attack
• 4.4.2.3 Coarse access
• 4.4.2.4 Physical copy attack
• 4.4.3 Design goals
• 4.5 Proposed scheme
• 4.5.1 Basic settings
• 4.5.2 Hierarchical key-role-area access-control model
• 4.5.3 Image publication
• 4.5.4 Client conformance
• 4.5.5 Case study
• 4.6 Security and performance analysis
• 4.6.1 Security analysis
• 4.6.2 Performance analysis
• 4.7 Conclusions
• Acknowledgments
• References

Inspec keywords: Big Data; authorisation; access control; cryptography; image watermarking; data analysis

Other keywords: geographic image; in situ access-control model; watermark-based access-control model; large-sized images; fine-grained access control; image big data; big data analysis; access-control policies; single image; hierarchical key-role-area model; image privacy

Subjects: Image and video coding; Data security; Other DBMS; Computer vision and image processing techniques; Cryptography