Traditional steganography method often hides secret data by establishing a mapping relationship between secret data and a cover image or directly in a noisy area, but has a low embedding capacity. Based on the thought of deep learning, in this work, we propose a new image steganography scheme based on a U-Net structure. First, in the form of paired training, the trained deep neural network includes a hiding network and an extraction network; then, the sender uses the hiding network to embed the secret image into another full-size image without any modification and send it to the receiver. Finally, the receiver uses the extraction network to reconstruct secret image and original cover image correctly. Experimental results show that the proposed scheme compresses and distributes the information of embedded secret image into all available bits in the cover image, which not only solves the obvious visual cues problem, but also increases the embedding capacity.

Image steganography is a technology that hides sensitive information into an image. The traditional image steganography method tends to securely embed secret information in the host image so that the payload capacity is almost ignored and the steganographic image quality needs to be improved for the Human Visual System(HVS). Therefore, in this work, we propose a new high capacity image steganography method based on deep learning. The Discrete Cosine Transform(DCT) is used to transform the secret image, and then the transformed image is encrypted by Elliptic Curve Cryptography(ECC) to improve the anti-detection property of the obtained image. To improve steganographic capacity, the SegNet Deep Neural Network with a set of Hiding and Extraction networks enables steganography and extraction of full-size images. The experimental results show that the method can effectively allocate each pixel in the image so that the relative capacity of steganography reaches 1. Besides, the image obtained using this steganography method has higher Peak Signal-to-Noise Ratio(PSNR) and Structural Similarity Index(SSIM) values, reaching 40dB and 0.96, respectively.

Image-to-image steganography is hiding one image in another image. However, hiding two secret images into one carrier image is a challenge today. The application of image steganography based on deep learning in real-life is relatively rare. In this paper, a new Steganography Convolution Neural Network (SteganoCNN) model is proposed, which solves the problem of two images embedded in a carrier image and can effectively reconstruct two secret images. SteganoCNN has two modules, an encoding network, and a decoding network, whereas the decoding network includes two extraction networks. First, the entire network is trained end-to-end, the encoding network automatically embeds the secret image into the carrier image, and the decoding network is used to reconstruct two different secret images. The experimental results show that the proposed steganography scheme has a maximum image payload capacity of 47.92 bits per pixel, and at the same time, it can effectively avoid the detection of steganalysis tools while keeping the stego-image undistorted. Meanwhile, StegaoCNN has good generalization capabilities and can realize the steganography of different data types, such as remote sensing images and aerial images.

Aiming at the problem that the traditional steganography based on carrier modification has the low steganographic capacity, a steganographic scheme based on Fully Convolutional Dense Connection Network (FC-DenseNet) is proposed. Since FC-DenseNet can effectively overcome the problems of gradient dissipation and gradient explosion, and a large number of features are multiplexed, the cascaded secret image and carrier image can reconstruct good image quality after entering the network. Effectively improve steganographic capacity. First, we reset the number of input channels of the first convolution block of FC-DenseNet and the number of output channels of the last convolution block and deleted the LogSoftmax() function. On the sender side, after the concatenated secret image and carrier image pass through the hidden network FC-DenseNet, the secret image is embedded in the carrier image to obtain a stego-image. At the receiving side, the extraction network reconstructs the secret image from the stego-image. Experimental results show that our proposed steganography scheme not only has a high Peak Signal-to-Noise Ratio(PSNR) and Structural Similarity(SSIM) but also can realize large-capacity image steganography, with an average image payload capacity of 23.96 bit per pixel.

The traditional cover modification steganography method only has low steganography ability. We propose a steganography method based on the convolutional neural network architecture (Xception) of deep separable convolutional layers in order to solve this problem. The Xception architecture is used for image steganography for the first time, which not only increases the width of the network, but also improves the adaptability of network expansion, and adds different receiving fields to carry out multi-scale information in it. By introducing jump connections, we solved the problems of gradient dissipation and gradient descent in the Xception architecture. After cascading the secret image and the mask image, high-quality images can be reconstructed through the network, which greatly improves the speed of steganography. When hiding, only the secret image and the cover image are cascaded, and then the secret image can be embedded in the cover image through the hidden network in order to obtain the secret image. After extraction, the secret image can be reconstructed by bypassing the secret image through the extraction network. The results show that the results that are obtained by our model have high peak signal-to-noise ratio (PSNR) and structural similarity (SSIM), and the average high load capacity is 23.96 bpp (bit per pixel), thus realizing large-capacity image steganography surgery.