Search

This article introduces a novel RNN unit inspired by GRU, namely the Content-Adaptive Recurrent Unit (CARU). The design of CARU contains all the features of GRU but requires fewer training parameters. We make use of the concept of weights in our design to analyze the transition of hidden states. At the same time, we also describe how the content adaptive gate handles the received words and alleviates the long-term dependence problem. As a result, the unit can improve the accuracy of the experiments, and the results show that CARU not only has better performance than GRU, but also produces faster training. Moreover, the proposed unit is general and can be applied to all RNN related neural network models.

This paper employs the concept of the content-adaptive residual and presents a low-dimensional feature set for detecting the grayscale steganography in spatial domain. The testing image is first segmented into three kinds of areas, that is, the smooth, edge, and textural areas. Then, different pixel predictors are used to calculate the residuals responded to different areas. The yielded different co-occurrence matrices are finally collected as the steganalytic features. Experiments reported show that the proposed method is effective and yields good performances when detecting popular steganographic algorithms such as LSB matching, EA, and HUGO.