Events Calendar

PhD Preliminary Oral Exam – Kimia Tajik

Balancing Image Privacy and Usability on the Cloud

In this work, we propose and analyze image encryption algorithms that balance privacy and usability on the cloud. First, we introduce ideal Thumbnail-Preserving Encryption (TPE), a special case of format-preserving encryption, and present a concrete construction. In ideal TPE, a ciphertext is itself an image that has the same thumbnail as the plaintext (unencrypted) image, but that provably leaks nothing about the plaintext beyond its thumbnail. We provide a formal security analysis for the construction, and a prototype implementation to demonstrate compatibility with existing photo storage services. We also study the ability of users to distinguish between thumbnail images preserved by TPE. Our findings indicate that TPE is an efficient and promising approach to balance usability and privacy concerns for images. In addition to leaving users' cloud usage pattern untouched, TPE does not require any modification to the cloud service backend. One direction of this problem that needs further investigation is the effectiveness of image super-resolution techniques to upscale the leaked pixelized image (thumbnail), reconstruct a higher resolution image and thus retrieve hidden information. Recent work on image privacy has shown that differentially private pixelization is achievable by adding Laplacian noise to the thumbnail. As future work, we propose leveraging similar noise with appropriate parameters as a defense against super-resolution. In order to benefit this new method in practice, we need a classic algorithm in the backend of cloud storage service, where encryption and decryption are implemented, and a well-sized thumbnail with appropriate noise in the front end, where users' interaction with the cloud service interface happens. Our proposed method does not modify users' familiar cloud usage pattern, as it is similar to the access pattern of TPE. However, achieving theoretical privacy (differential privacy) ! and practical privacy (resistance to super-resolution) in addition to usability, comes at the cost of requiring support from cloud provider.

Major Advisor: Rakesh Bobba
Committee: Mike Rosulek
Committee: Fuxin Li
Committee: Charles Wright
GCR: Yelda Turkan

Tuesday, March 17 at 1:00pm to 3:00pm

Kelley Engineering Center, 1005
110 SW Park Terrace, Corvallis, OR 97331

Event Type

Lecture or Presentation

Event Topic

Research

Organization
Electrical Engineering and Computer Science
Contact Name

Dakota Nelson

Contact Email

eecs.gradinfo@oregonstate.edu

Subscribe
Google Calendar iCal Outlook

Recent Activity