Research on Steganalysis of Digital Image Based on Deep Learning

Steganography is a technique and science of information hiding. Corresponding to the development of image steganography, steganalysis has made substantial progress. In recent years, deep learning is widely used in the fields of computer vision and natural language processing. With people's in-depth understanding of deep learning, convolutional neural network technology has been introduced into the field of steganography and steganalysis, and has made a series of breakthroughs and development in methods and performance. Under this background, this paper summarizes and discusses the main methods and the most representative work of how to apply convolutional neural network model of deep learning to steganalysis of digital images. Finally, it discusses the problems and future development of steganalysis and steganalysis of digital image based on deep learning.