Adverse weather conditions, such as rain, raindrop, snow and haze, consistently degrade images in an unpredictable manner, thereby rendering existing task-specific and task-aligned methods inadequate in addressing this formidable problem. To this end, we investigate the application of Transformer in image restoration and introduce an efficient frequency-oriented method called AIRFormer, which is designed to restore weather-degraded images comprehensively and holistically. Specifically, we identify that the initial self-attention mechanism exhibits distinctive properties akin to a low-pass filter. Therefore, we construct a frequency-guided Transformer encoder by incorporating wavelet-based prior information to guide the extraction of image features. Additionally, considering the non-specific frequency characteristics of self-attention in the later stages, we develop a frequency-refined Transformer decoder that incorporates learnable task-specific queries across spatial dimensions, channel dimensions, and wavelet domains. To facilitate the training of our proposed method, we curate a comprehensive benchmark dataset named AIR40K that, encompasses a wide range of challenging scenarios. Extensive experimental evaluations demonstrate the superiority of our AIRFormer over both task-aligned and all-in-one methods across 15 publicly available datasets. Notably, AIRFormer achieves the best trade-off between the inference time and quality of reconstructed image, comparing with existing methods such as TransWeather and Restormer. The source code, dataset and pre-trained models will be available at https://github.com/chdwyb/AIRFormer.

In this paper, a secure and fast chaotic image encryption scheme with simultaneous permutation-diffusion operation is proposed. We combine permutation and diffusion processes into a whole, namely, simultaneous permutation and diffusion operation (SPDO). This can solve the problem of traditional encryption scheme in which the permutation and diffusion are two independent processes, that leads attackers to crack the two processes separately. In SPDO, the initial value of the current Sine-Sine chaotic map is related to the secret keys and the previous encrypted pixels’ values. In this case, the proposed scheme can generate dynamic key streams and indexes that are related to plaintext, which improves the sensitivity to plaintext for the encryption scheme. In addition, the pixel values are processed by row and column (row-level and column-level) during the encryption procedure. Thus, the proposed scheme presents lower time complexity and faster running speed compared with bit-level or pixel-level image encryption schemes, which makes the proposed scheme be conducive to the batch transmission and real-time transmission of digital images. The simulation results and security analysis show that our scheme can resist common attacks, such as statistical attack, differential attack, chosen plaintext attack and other comprehensive attacks.

The removal of rain streaks and raindrops is crucial for enhancing the image visibility and mitigating the weather degradations. However, most existing approaches rely on the paired rainy and clean images, which are challenging to obtain in real-world scenarios. To this end, we propose a novel structure-preserving frequency-regularized text-guided optimal transport (SFTOT) framework, which formulates the unpaired rain streaks and raindrops removal as an optimal transport problem. Specifically, we introduce a structure-preserving transport cost, incorporating the structural similarity constraint to minimize the duality gap between the primal and dual formulations, while preserving the structural details of reconstructed images. Furthermore, by embedding the inherent frequency sparsity of rain streaks and raindrops into the transport cost, we derive a frequency-regularized optimal transport objective, ensuring consistency in frequency distributions between the generated and clean images. Additionally, we employ a pre-trained one-step stable diffusion model as the restoration network, which is fine-tuned using the low-rank adaptation (LoRA) adapters and zero convolutional layers, while integrating the domain-specific text prompts for both degraded and clean images to guide the generation process. Extensive experiments demonstrate that our method surpasses the existing well-performing unpaired learning approaches, achieving notable improvements in both the fidelity and photo-realism.

The classic scrambling-diffusion structure for image encryption is vulnerable to be attacked. The scrambling and diffusion process can be cracked separately due to the two processes are independent. In this paper, we proposed a novel color image encryption algorithm which can overcome the problem. In the proposed algorithm, firstly, the first diffusion is performed simultaneously with the first scrambling which can resist the attacker to separately crack the scrambling or diffusion. Then, the second diffusion matrix generated from the Hopfield chaotic neural network (HCNN) which is related to the plaintext is performed which can not only resist common chosen-plaintext attack, but also improved the key sensitivity. Finally, a second scrambling is performed on some special pixels to further improve the security. Experiments and security analysis indicate that the proposed algorithm has an excellent performance in image encryption and various attacks.

Existing approaches have significantly advanced remote sensing image dehazing. However, they often rely on conventional encoder-decoder architectures, leading to prolonged inference times. To this end, we propose a novel cross-level interaction and intra-level fusion network for remote sensing image dehazing (CINet), which shifts the focus from encoder-decoder dependencies to an innovative hierarchical architecture centered on skip connections, leading to competitive dehazing performance with decreased calculating complexity. Furthermore, we introduce a cross-level multi-view interaction module (CMIM) to facilitate effective interactions between features across hierarchical levels, mitigating the information loss commonly caused by repeated down-sampling operations. Meanwhile, we develop an intra-level dual-dimension fusion module (IDFM), which leverages height-wise and width-wise self-attention to capture rich spatial-aware information, enabling robust and efficient intra-level feature fusion. Additionally, we propose a multi-view progressive extraction block (MPEB), which decomposes features into four distinct components and applies convolutions with diverse kernel sizes, groups, and dilation factors. This design promotes progressive feature learning while significantly reducing computational overhead. Extensive experiments conducted on nine publicly available datasets validate the effectiveness and superiority of our proposed model.

In this paper, a novel double-image compression-encryption algorithm is proposed by combining Rossler hyperchaos and compressive sensing. In the proposed scheme, the sparse matrixes obtained by performing two-dimensional discrete wavelet transform on two plain images are first scrambled and compressed by index confusion and compressive sensing. Next, the compressed matrixes are linearly quantized, and then variable step length Josephus confusion and bitwise exclusive-OR operation are performed on them. The final noise-like cipher image is generated by hiding half of the encrypted image into the alpha channel of the other half of encrypted image. In order to realize “One cipher image corresponds to one key”, a novel key generation mechanism is designed. Finally, the simulation analyses indicate that the proposed double-image encryption scheme has high transmission efficiency. Meanwhile, the average local information entropy of the cipher image is about 0.902 and the decryption quality is as high as 30 dB.