Timezone: »

Fixed Neural Network Steganography: Train the images, not the network
Varsha Kishore · Xiangyu Chen · Yan Wang · Boyi Li · Kilian Weinberger
Event URL: https://eventhosts.gather.town/app/kR7ip0Bhhn8BXuMD/wiml-workshop-2021 »

Recent attempts at image steganography make use of advances in deep learning to train an encoder-decoder network pair to hide and retrieve secret messages in images. These methods are able to hide large amounts of data, but also incur high decoding error rates (around 20\%). We propose a novel algorithm for steganography that takes advantage of the fact that neural networks are sensitive to tiny perturbations. Our method, Fixed Neural Network Steganography (FNNS), achieves 0\% error reliably for hiding up to 3 bits per pixel (bpp) of secret information in images and yields significantly lower error rates when compared to prior state of the art methods for hiding more than 3 bpp. FNNS also successfully evades existing statistical steganalysis systems and can be modified to evade neural steganalysis systems as well. Recovering every bit correctly, up to 3 bpp, enables novel applications, e.g. those requiring encryption. We introduce one specific use case for facilitating anonymized and safe image sharing.

Author Information

Varsha Kishore (Cornell University)
Xiangyu Chen (Cornell University)
Yan Wang (Cornell)
Boyi Li (Cornell University)
Kilian Weinberger (Cornell University / ASAPP Research)

More from the Same Authors