No abstract

Key distribution is the foundation for protecting users' privacy and communication security in cloud environment. Information hiding is an effective manner to hide the transmission behavior of secret information such as keys, and thus it makes the secure key distribution possible. However, the traditional information hiding systems usually embed the secret information by modifying the carrier, which inevitably leaves modification traces on the carrier. Thus, they cannot resist the detection of the steganalysis algorithm effectively. To avoid this issue, the coverless information hiding technique has been proposed accordingly, in which the original images of which features can express the secret information are directly used as stegoimages. Since the existing coverless information hiding methods use the low-level handcrafted image features to express secret information, it is hard for them to realize desirable robustness against common image attacks. Moreover, their hiding capacity is limited. To conquer these problems, we design a novel robust image coverless information hiding system using Faster Region-based Convolutional Neural Networks (Faster-RCNN). We employ Faster-RCNN to detect and locate objects in images and utilize the labels of these objects to express secret information. Since the original images without any modification are used as stego-images, the proposed method can effectively resist steganalysis and will not cause attackers' suspicion. The experimental results demonstrate that the proposed system has higher performance in terms of robustness and capacity compared to the typical coverless information hiding methods.

The well-known Thien and Lin's (k, n) secret image sharing (SIS) scheme and its extended versions are threshold schemes, in which a secret image is shared among n shadow images and it can be recovered from any k shadow images. To reduce the size of shadow image, in those schemes, secret image pixels are embedded in all coefficients of (k −1)-degree polynomial to generate the shadows. Also, the secret pixels are permuted before the sharing to address the residual-image problem on shadow images. Due to the above two approaches, partial secret information can be exposed from (k − 1) shadow images, and thus the threshold properties of those schemes will be compromised. To overcome this weakness, we propose a novel (k, n)-SIS scheme based on encrypted pixels, whose shadow image size is slightly larger than that of Thien and Lin's scheme. By slightly modifying the secret image, we also propose a modified (k, n)-SIS scheme with the same shadow size of Thien and Lin's scheme.

To fundamentally resist the steganalysis, coverless information hiding has been proposed, and it has become a research hotspot in the field of covert communication. However, the current methods not only require a huge image database, but also have a very low hidden capacity, making it difficult to apply practically. In order to solve the above problems, we propose a coverless information hiding method based on the generation of anime characters, which first converts the secret information into an attribute label set of the anime characters, and then uses the label set as a driver to directly generate anime characters by using the generative adversarial networks (GANs). The experimental results show that compared with the current methods, the hidden capacity of the proposed method is improved by nearly 60 times, and it also has good performance in image quality and robustness.

Due to the lack of pre-judgment of fingerprints, fingerprint authentication systems are frequently vulnerable to artificial replicas. Anonymous people can impersonate authorized users to complete various authentication operations, thereby disrupting the order of life and causing tremendous economic losses to society. Therefore, to ensure that authorized users' fingerprint information is not used illegally, one possible anti-spoofing technique, called fingerprint liveness detection (FLD), has been exploited. Compared with the hand-crafted feature methods, the deep convolutional neural network (DCNN) can automatically learn the high-level semantic detail via supervised learning algorithm without any professional background knowledge. However, one disadvantage of most CNNs models is that fixed scale images (e.g., 227 × 227) are essential in the input layer. Although the scale problem can be handled by cropping or scaling operations via transforming an image of any scale into a fixed scale, they can easily cause some key texture information loss and image resolution degradation, which will weaken the generalization performance of the classifier model. In this paper, a novel FLD method called an improved DCNN with image scale equalization, has been proposed to preserve texture information and maintain image resolution. Besides, an adaptive learning rate method has been used in this paper. In the performance evaluation, the confusion matrix is applied into FLD for the first time as a performance indicator. The amounts of the experimental results based on the LivDet 2011 and LivDet 2013 data sets also verify that the detection performance of our method is superior to other methods.