P3: Toward Privacy-Preserving Photo Sharing
summary
P3 extracts and encrypts the image’s most informative DCT components (extracts and encrypts DC and large AC DCT coefficients), leaves the rest as a standards‑compatible JPEG (public part remains standard JPEG), and stores them separately so providers can still do server‑side scaling/cropping. Prototype integrates with Facebook, runs on a mobile proxy, uses an operating threshold range 1–20, adds ≈20% worst‑case storage over originals, evaluates on 1,491 images, and reduces face‑recognition success to about 2% at threshold 1 while enabling correct reconstruction after linear cloud transforms.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 0:22Privacy Concerns
- 4:10Framework Assumptions
- 5:13Framework Approach
- 7:13Intuition
- 12:34De decryption
- 14:17Design
- 17:26Results
related talks
40mNew Face, Who Dis? Protecting Privacy in a World of Surveillance
11mPrivacy-Preserving Protocols for Smart Camera Systems and Other IoT Devices
17mEfficient & Frugal Image Data Processing Pipelines
03mDelivering fast and beautiful images and video
33mMastering the Images Spectrum
14mEstablishing Privacy Metrics for Genomic Data Analysis
21mCombining Privacy and AI
01h 11mMoving Pictures