Steganographic Embeddings as an Effective Data Augmentation

TL;DR

This paper demonstrates that LSB steganography can be effectively used as a data augmentation technique in deep learning, improving training efficiency and approximating color augmentations without extra hyperparameter tuning.

Contribution

It introduces a novel application of steganography as a data augmentation method for deep learning, showing its benefits in training efficiency and augmentation approximation.

Findings

LSB steganography enhances deep neural network training on CIFAR-10.

It acts as a discretized approximation of color augmentations.

No additional training overhead or hyperparameter tuning is required.

Abstract

Image Steganography is a cryptographic technique that embeds secret information into an image, ensuring the hidden data remains undetectable to the human eye while preserving the image's original visual integrity. Least Significant Bit (LSB) Steganography achieves this by replacing the k least significant bits of an image with the k most significant bits of a secret image, maintaining the appearance of the original image while simultaneously encoding the essential elements of the hidden data. In this work, we shift away from conventional applications of steganography in deep learning and explore its potential from a new angle. We present experimental results on CIFAR-10 showing that LSB Steganography, when used as a data augmentation strategy for downstream computer vision tasks such as image classification, can significantly improve the training efficiency of deep neural networks. It can also act as an implicit, uniformly discretized piecewise linear approximation of color augmentations such as (brightness, contrast, hue, and saturation), without introducing additional training overhead through a new joint image training regime that disregards the need for tuning sensitive augmentation hyperparameters.