Section: New Results

Privacy

More and more users prefer to share their photos through image-sharing platforms of social networks than using e-mail or personal webpages. Since the provider of the image-sharing platform can clearly know the contents of any published images, the users have to trust the provider to respect their privacy or has to encrypt their images. In the context of the PhD of Kun He [18], [17], [16], we have proposed an IND-CPA image encryption algorithm that preserve the image format after encryption, and we have shown that our encryption algorithm can be used on several widely used image-sharing platforms such as Flickr, Pinterest, Google+ and Twitter.

Active fingerprinting schemes were originally invented to deter malicious users from illegally releasing an item, such as a movie or an image. To achieve this, each time an item is released, a different fingerprint is embedded in it. In the context of the PhD of Julien Lolive, we have defined the first privacy-preserving asymmetric fingerprinting protocol based on Tardos codes [2]. This protocol is optimal with respect to traitor tracing. We also formally proved that our protocol achieves the properties of correctness, anti-framing, traitor tracing, as well as buyer- and item-unlinkability.