In this paper, we present the multi-stage attentive network (MSAN), an efficient and good generalization performance convolutional neural network (CNN) architecture for motion deblurring. We build a multi-stage encoder–decoder network with self-attention and use the binary cross-entropy loss to train our model. In MSAN, there are two core designs. First, we introduce a new attention-based end-to-end method on top of multi-stage networks, which applies group convolution to the self-attention module, effectively reducing the computing cost and improving the model’s adaptability to different blurred images. Secondly, we propose using binary cross-entropy loss instead of pixel loss to optimize our model to minimize the over-smoothing impact of pixel loss while maintaining a good deblurring effect. We conduct extensive experiments on several deblurring datasets to evaluate the performance of our solution for deblurring. Our MSAN achieves superior performance while also generalizing and compares well with state-of-the-art methods.

Under the framework of computational integral imaging, a multi-image encryption scheme based on the DNA-chaos algorithm is proposed. In this scheme, multiple images are merged to one image by a computational integral imaging algorithm, which significantly improves the efficiency of image encryption. Meanwhile, the computational integral imaging algorithm can merge images at different depth distances, thereby the different depth distances of multiple images can also be used as keys to increase the security of the encryption method. In addition, the high randomness of the chaos algorithm is combined to address the outline effect caused by the DNA encryption algorithm. We have experimentally verified the proposed multi-image encryption scheme. The entropy value of the encrypted image is $7.6227$, whereas the entropy value of the merge image with two input images is $3.2886$, which greatly reduces the relevance of the image. The simulation results also confirm that the proposed encryption scheme has high key security and can protect against various attacks.

A novel adaptive secure holographic image watermarking method in the sharp frequency localized contourlet transform (SFLCT) domain is presented. Based upon the sine logistic modulation map and the logistic map, we develop an encrypted binary computer-generated hologram technique to fabricate a hologram of a watermark first. Owing to the enormous key space of the encrypted hologram, the security of the image watermarking system is increased. Then the hologram watermark is embedded into the SFLCT coefficients with Schur decomposition. To obtain better imperceptibility and robustness, the entropy and the edge entropy are utilized to select the suitable watermark embedding positions adaptively. Compared with other watermarking schemes, the suggested method provides a better performance with respect to both imperceptibility and robustness. Experiments show that our watermarking scheme for images is not only is secure and invisible, but also has a stronger robustness against different kinds of attack.

In this paper, a watermark embedding scheme based on ghost image encryption and cellular automata transformation is proposed. In this scheme, the watermark forms speckle through different light intensities into a key, and the cellular automata transformation algorithm is embedded into the 3D image. Compared with the traditional watermarking encryption method, this scheme combines ghost imaging and the cellular automata transformation algorithm, which double guarantees and increases the confidentiality of the watermark. The binary computing ghost imaging discussed in this paper saves the storage space of password text and makes the transmission of password text more convenient and faster. Experiments on this method also verify that the watermark-embedded image has higher imperceptibility and higher robustness against attacks, and that the extracted watermark has good integrity.