Security and Privacy of Visual Data

1
Security and Privacy of Visual Data

• Maneesh Upmanyu, C. Narsimha RajuAnoop M.
  Namboodiri, K. Srinathan, C.V. Jawahar
• Center for Visual Information Technology
• Center for Security, Theory Algorithmic
  Research
• IIIT- Hyderabad

2
Broad Objective

• Development of secure computational algorithms in
  computer vision and related areas.
• To develop
• Provably secure solutions
• computationally efficient solutions
• solutions to problems with immediate impact

Blind Vision (Avidan and Butman, ECCV 2006)
raised and addressed the privacy concerns in
vision.
3
Issues to be addressed

• Efficiency Encryption/Decryption is
  computationally expensive. Secure algorithms
  should be practical.
• Domain Knowledge Domain specific algorithms will
  be more efficient than generic solutions such as
  SMC, AES.
• Security Algorithms need to be provably-secure
  and meet futuristic requirements.

4
Research Directions

• Video Protection using various Encryption
  Algorithms
• Private Content Based Image Retrieval (PCBIR)
• Blind Authentication A Secure Crypto-Biometric
  Verification Protocol
• Efficient Privacy Preserving Video Surveillance

5
Video Encryption

• Video protection using encryption is an
  extremely useful method for the stopping unwanted
  interception and viewing of any transmitted video
  or other information.
• Secure encryption algorithms designed to address
  this. We exploit the properties of the DCT
  coefficients and their influence on the visual
  content for designing this encryption scheme and
  their statistical behavior.
• Proposed algorithm is secure and make video
  encryption possible for real-time video
  applications.
• Publication C. Narsimha Raju, Umadevi Ganugula,
  K. Srinathan and C.V. Jawahar Fast and Secure
  Real-time Video Encryption in Proceedings of
  the 6th Indian Conference on Computer Vision,
  Graphics and Image Processing (ICVGIP-09)

6
Private Content Based Image Retrieval (PCBIR)

• Problem setting Alice has query image (I), Bob
  maintains a image-database server. Alice wants to
  retrieve similar images without revealing (I) to
  Bob.
• Secure algorithms designed to address this.
  Exploits the clustered nature of image databases
  to reduce the amount of communication required in
  SMC
• Proposed algorithms are completely private and
  feasible for extremely large datasets

• Publication Shashank J, Kowshik P, Kannan
  Srinathan and C.V. Jawahar Private Content Based
  Image Retrieval Computer Vision and Pattern
  Recognition(CVPR 2008)

7
Blind Authentication A Secure Crypto-Biometric
Verification Protocol

• First ever solution to completely address the
  concerns of users privacy, template protection,
  and trust issues in biometric authentication
  systems
• Problem setting Alice wants to create an account
  in Bob-mail, that requires biometrics based
  authentication. However, she neither trusts Bob
  to handle her biometric data securely, nor trusts
  the network to send her plain biometric
• Captures advantages of biometrics as well as
  security of public key cryptography

• Publication Maneesh Upmanyu, Anoop M.
  Namboodiri, K. Srinathan and C.V. Jawahar
  Efficient Biometric Verification in Encrypted
  Domain In Proceedings of the 3rd International
  Conference on Biometrics (ICB 2009)

8
Efficient Privacy Preserving Video Surveillance

• Problem Setting Alice wants to have her office
  environment under surveillance by Bob. If Alice
  wants to track a person Jack, then Bob should
  have access to track Jack and no one else.
• Utilize some of the inherent properties of image
  data to enable efficient and provably secure
  surveillance algorithms
• Provably secure, real-time detection and tracking
  of people

• Publication Maneesh Upmanyu, Anoop M.
  Namboodiri, K. Srinathan and C.V. Jawahar
  Efficient Privacy Preserving Video Surveillance.
  In Proceedings of the 12th International
  Conference on Computer Vision (ICCV 2009)

9
Results of Video Encryption
10
Example 1 Presentation