Review on Image Steganalysis Using INRIA Dataset

Abstract

Steganography can be defined as the science of private communication while steganalysis is the science of detection of embedded messages in digital media with the use of steganography (Each of steganography and steganalysis are under much interest from media and law enforcement entities). Blind steganalysis had the attempt of differentiating steganographic images from cover images without knowing the method of steganography.

Popular data-sets are common in various areas like image processing, artificial intelligent, and security which allow the researchers to validate their approaches. INRIA dataset is one type of image dataset which include Holiday, Copydays and BIGANN evaluation datasets. This paper presents study on number of researches using INRIA dataset for image/ information retrieval and especially blind image steganalysis. All these works depends on statistical properties of image. No one use machine learning tools like deep learning especially convolution neural network to detect attack in image using INRIA dataset.