Defeating Image Obfuscation with Deep Learning

TL;DR

This paper shows that deep learning models can effectively recover recognizable information from obfuscated images, including mosaicing, blurring, and encrypted JPEG coefficients, challenging privacy-preserving techniques.

Contribution

It demonstrates that neural networks can break various image obfuscation methods, revealing vulnerabilities in current privacy-preserving image sharing systems.

Findings

Neural networks can identify faces in mosaiced images.

Deep learning can recognize objects in blurred images.

Models successfully classify handwritten digits from encrypted JPEG images.

Abstract

We demonstrate that modern image recognition methods based on artificial neural networks can recover hidden information from images protected by various forms of obfuscation. The obfuscation techniques considered in this paper are mosaicing (also known as pixelation), blurring (as used by YouTube), and P3, a recently proposed system for privacy-preserving photo sharing that encrypts the significant JPEG coefficients to make images unrecognizable by humans. We empirically show how to train artificial neural networks to successfully identify faces and recognize objects and handwritten digits even if the images are protected using any of the above obfuscation techniques.