Pansharpening is a powerful technique for generating high-resolution multispectral (HRMS) images by fusing currently available image pairs of low-resolution multispectral (LRMS) and texture-rich panchromatic (PAN) data, effectively addressing the physical constraints of satellite sensors. While recent generative diffusion models have demonstrated impressive performance gains in this domain, their prohibitive computational demands and training costs hinder practicality in resource-constrained remote sensing satellite systems. In this work, we propose NODiff, a novel diffusion framework that replaces the conventional attention-based denoising backbone with a neural operator, seamlessly integrating operator learning and generative modeling into an efficient yet effective solution for pansharpening. In practice, we implement our approach through a two-stage learning paradigm: First, we pretrain the proposed Neural Operator-based diffusion model to learn the high-resolution texture priors essential for pansharpening. Afterward, we freeze the pretrained parameters, and design a lightweight conditional detail guidance adapter to enable efficient fine-tuning for generating desired HRMS images. Meanwhile, a time-aware low-rank adaptation is introduced to dynamically refine high-frequency details potentially affected by spectral mode truncation. Extensive experiments on multiple benchmark datasets demonstrate that NODiff achieves competitive pansharpening performance while significantly reducing training and inference costs. Beyond pansharpening, our method provides new insights into building resource-efficient generative models.

Fully fine-tuning large pre-trained models for each downstream task is impractical due to prohibitive memory, computation, and storage costs. Although parameter-efficient fine-tuning (PEFT) methods address this issue, leading methods like LoRA still exhibit linear scaling of trainable parameters with hidden size. Recent studies have explored PEFT in the frequency domain to reduce computational costs by employing fast Fourier transform and discrete cosine transform with sparse frequency selection. These methods rely on global frequency representations that lack spatial locality and disperse energy across the domain. As a result, sparse coefficient selection struggles to preserve fine-grained structural information and often introduces artifacts such as ringing near boundaries. To address these limitations, we propose DWTSG, a novel PEFT framework based on discrete wavelet transform (DWT) and subband guidance. DWTSG decomposes pre-trained weights into four wavelet subbands that jointly encode global context and local details. It fine-tunes only the most informative coefficients in each subband through an energy-based selection strategy that prioritizes coefficients based on their individual importance and interactions. Finally, inverse DWT reconstructs the updated weights, enabling efficient and precise adaptation. Extensive experiments on natural language understanding, commonsense reasoning, and image classification demonstrate that DWTSG outperforms existing PEFT methods, achieving superior performance and higher parameter efficiency.

The rapid growth of model scale has necessitated substantial computational resources for fine-tuning. Existing approach such as low-rank adaptation (LoRA) has sought to address the problem of handling the large updated parameters in full fine-tuning (FT). However, LoRA utilize random initialization and optimization of low-rank matrices to approximate updated weights, which can result in suboptimal convergence and an accuracy gap compared to full fine-tuning (FT). To address these issues, we propose low-rank LDU (LoLDU), a parameter-efficient fine-tuning (PEFT) approach that significantly reduces trainable parameters by 2600 times compared to regular PEFT methods while maintaining comparable performance. LoLDU leverages lower-diag-upper (LDU) decomposition to initialize low-rank matrices for faster convergence and nonsingularity. We focus on optimizing the diagonal matrix for scaling transformations. To the best of our knowledge, LoLDU has the fewest parameters among all PEFT approaches. We conducted extensive experiments across 4 instruction-following datasets, six natural language understanding (NLU) datasets, eight image classification datasets, and image generation datasets with multiple model types [LLaMA2, RoBERTa, ViT, and stable diffusion (SD)], providing a comprehensive and detailed analysis. Our open-source code can be accessed at https://anonymous.4open.science/r/LoLDU-B5A6.

Video Moment Retrieval (VMR) involves locating specific moments within a video based on natural language queries. However, existing VMR methods that employ various strategies for cross-modal alignment still face challenges such as limited understanding of fine-grained semantics, semantic overlap, and sparse constraints. To address these limitations, we propose a novel Concept Decomposition Transformer (CDTR) model for VMR. CDTR introduces a semantic concept decomposition module that disentangles video moments and sentence queries into concept representations, reflecting the relevance between various concepts and capturing fine-grained semantics which is crucial for cross-modal matching. These decomposed concept representations are then used as pseudo-labels, determined as positive or negative samples by adaptive concept-specific thresholds. Subsequently, fine-grained concept alignment is performed in video intra-modal and textual-visual cross-modal, aligning different conceptual components within features, enhancing the model's ability to distinguish fine-grained semantics, and alleviating issues related to semantic overlap and sparse constraints. Comprehensive experiments demonstrate the effectiveness of the CDTR, outperforming state-of-the-art methods on three widely used datasets: QVHighlights, Charades-STA, and TACoS.

Pan-sharpening technology refers to generating a high-resolution (HR) multi-spectral (MS) image with broad applications by fusing a low-resolution (LR) MS image and HR panchromatic (PAN) image. While deep learning approaches have shown impressive performance in pan-sharpening, they generally require extensive hardware with high memory and computational power, limiting their deployment on resource-constrained satellites. In this study, we investigate the use of binary neural networks (BNNs) for pan-sharpening and observe that binarization leads to distinct information degradation across different frequency components of an image. Building on this insight, we propose a novel binary pan-sharpening network, termed BNNPan, structured around the Prior-Integrated Binary Frequency (PIBF) module that features three key ingredients: Binary Wavelet Transform Convolution, Latent Diffusion Prior Compensation, and Channel-wise Distribution Calibration. Specifically, the first decomposes input features into distinct frequency components using Wavelet Transform, then applies a "divide-and-conquer" strategy to optimize binary feature learning for each component, informed by the corresponding full-precision residual statistics. The second integrates a latent diffusion prior to compensate for compromised information during binarization, while the third performs channel-wise calibration to further refine feature representation. Our BNNPan, developed with the proposed techniques, achieves promising pan-sharpening performance on multiple remote sensing datasets, surpassing state-of-the-art binarization algorithms.

Video grounding tasks have recently gained significant attention. However, existing methods failed to fully comprehend the semantics within queries and videos, often overlooking key content. Moreover, the lack of fine-grained cross-modal alignment and interaction to guide the semantic matching of complex texts and videos lead to inconsistent representational modeling. To address this issue, we propose a Semantic Hierarchical Grounding model, referred to as SHG, and design a cross-modal semantic hierarchical graph to achieve fine-grained semantic understanding. SHG decomposes both the query and each video moment into three levels: global, action, and element. This topology, ranging from global to local, establishes multigranularity intrinsic connections between the two modalities, fostering a comprehensive understanding of dynamic semantics and fine-grained cross-modal matching. Accordingly, to fully leverage the rich information within the cross-modal semantic hierarchical graph, we employ contrastive learning by seeking samples with the same action and element semantics, then achieve node-moment cross-modal hierarchical matching for global alignment. This approach can unearth fine-grained clues and align semantics across multiple granularities. Moreover, we combine the designed hierarchical graph interaction for coarse-to-fine fusion of text and video, thereby enabling highly accurate video grounding. Extensive experiments conducted on three challenging public datasets (ActivityNet-Captions, TACoS, and Charades-STA) demonstrate that the proposed approach outperforms state-of-the-art techniques, validating its effectiveness.

Temporal Sentence Grounding (TSG) requires a thorough understanding of the complex cross-modal semantic relationships between videos and text. However, existing methods fail to accurately capture content at diverse granularity levels with distinct semantics, making it difficult to achieve fine alignment of visuals and text. To overcome this issue, we attempt to mine for rich semantic clues by utilizing the hierarchical correspondence structure and multi-granularity visual-to-text reconstruction, achieving fine-grained reasoning. Specifically, for the TSG task, we propose a novel Hierarchical Cross-modal Fine-grained Mining Network (HCFMN), which utilizes an attention mechanism based on temporal hierarchical relationships to extract temporal features corresponding to the text of different granularities. We leverage the reconstructability of visual-to-text, recovering multi-granularity textual content from coarse to fine by focusing on temporal features at different layers, hierarchically extracting temporal features and the dependencies related to the text, and implementing fine-grained cross-modal semantic alignment. Furthermore, HCFMN introduces a novel partitioned efficient attention mechanism, which significantly enhances the model’s efficiency through a two-stage attention based on sequence and channel compression. Extensive experimental results on three public datasets (ActivityNet-Captions, TACoS, and Charades-STA) demonstrate that the proposed method achieves state-of-the-art performance.

Pixel-level adaptive convolution, which overcomes the deficiency of the spatial-invariance of standard convolution, is always limited to performing feature extraction from local patches and ignores the latent long-range dependencies imperceptible in the feature space, which are more significant in pixel-level tasks such as hyperspectral image super-resolution (HSISR). To handle such limitations, we propose kernel-space non-local convolution (KNLConv), which explores non-local dependencies in the generated kernel space, to leverage these global information to guide the network to extract image features more flexibly. Technically, the proposed KNLConv first decomposes the convolutional kernel space into spatial and channel dimensions, and designs a depth-wise non-local expansion convolution (NLEC) in the spatial dimension of the kernel-space to explore underlying global correlations. Then introduce an adaptive point-wise convolution (APC), generalizing the NLEC to the pixel-level while integrating features in the channel dimension. In addition, applying KNLConv, we design an effective network architecture for hyperspectral image super-resolution. Extensive experiments demonstrate that our approach performs favorably against current state-of-the-art HSISR methods, both on quantitative indicators and visual quality.

The vulnerability of deep neural networks to adversarial examples has raised huge concerns about the security of these algorithms. Black-box adversarial attacks have received a lot of attention as an influential method for evaluating model robustness. While various sophisticated adversarial attack methods have been proposed, the success rate in the black-box scenario still needs to be improved. To address these issues, we develop an Adaptive Multi-scale Degradation-based Attack method calledAMDA. The intuitive motivation behind our approach is that different models tend to have similar attention regions for low-scale images. Specifically, AMDA uses degraded images to generate perturbations at different scales and fuses these perturbations to generate adversarial examples that are insensitive to model changes. Furthermore, we design an adaptive multi-scale perturbation fusion that evaluates the transferability of perturbations at different scales based on noise and adaptively allocates fusion weights to prioritize strong transferability attacks and avoid being compromised by local optima. Extensive experimental results on the ImageNet, CIFAR-100, and CIFAR-10 datasets demonstrate that the proposed AMDA algorithm exhibits competitive performance for both normally trained models and defense models.

Transformer-based methods have proven to be effective in achieving long-distance modeling, capturing the spatial and spectral information, and exhibiting strong inductive bias in various computer vision tasks. Generally, the Transformer model includes two common modes of multi-head self-attention (MSA): spatial MSA (Spa-MSA) and spectral MSA (Spe-MSA). However, Spa-MSA is computationally efficient but limits the global spatial response within a local window. On the other hand, Spe-MSA can calculate channel self-attention to accommodate high-resolution images, but it disregards the crucial local information that is essential for low-level vision tasks. In this study, we propose a bidirectional dilation Transformer (BDT) for multispectral and hyperspectral image fusion (MHIF), which aims to leverage the advantages of both MSA and the latent multiscale information specific to MHIF tasks. The BDT consists of two designed modules: the dilation Spa-MSA (D-Spa), which dynamically expands the spatial receptive field through a given hollow strategy, and the grouped Spe-MSA (G-Spe), which extracts latent features within the feature map and learns local data behavior. Additionally, to fully exploit the multiscale information from both inputs with different spatial resolutions, we employ a bidirectional hierarchy strategy in the BDT, resulting in improved performance. Finally, extensive experiments on two commonly used datasets, CAVE and Harvard, demonstrate the superiority of BDT both visually and quantitatively. Furthermore, the related code will be available at the GitHub page of the authors.

Pansharpening is a crucial and challenging task that aims to obtain a high spatial resolution image by merging a multispectral (MS) image and a panchromatic (PAN) image. Current methods use CNNs with standard convolution, but we've observed strong correlation among channel dimensions in the kernel, leading to computational burden and redundancy. To address this, we propose Learnable Gaussian Perturbation Convolution (LGPConv), surpassing standard convolution. LGPConv leverages two properties of standard convolution kernels: 1) correlations within channels, learning a premier kernel as a base to reduce parameters and training difficulties caused by redundancy; 2) introducing Gaussian noise perturbations to simulate randomness and enhance nonlinear representation within channels. We incorporate LGPConv into a well-designed pansharpening network and demonstrate its superiority through extensive experiments, achieving state-of-the-art performance with minimal parameters (27K). Code is available on the GitHub page of the authors.

Pansharpening is a challenging low-level vision task whose aim is to learn the complementary representation between spectral information and spatial detail. Despite the remarkable progress, existing deep neural network (DNN) based pansharpening algorithms are still confronted with common limitations. 1) These methods rarely consider the local specificity of different spectral bands; 2) They often extract the global detail in the spatial domain, which ignore the task-related degradation, e.g., the down-sampling process of MS image, and also suffer from limited receptive field. In this work, we propose a novel bidomain modeling paradigm for pansharpening problem (dubbed as BiMPan), which takes into both local spectral specificity and global spatial detail. More specifically, we first customize the specialized source-discriminative adaptive convolution (SDAConv) for every spectral band instead of sharing the identical kernels across all bands like prior works. Then, we devise a novel Fourier global modeling module (FGMM), which is capable of embracing global information while benefiting the disentanglement of image degradation. By integrating the band-aware local feature and Fourier global detail from these two functional designs, we can fuse a texture-rich while visually pleasing high-resolution MS image. Extensive experiments demonstrate that the proposed framework achieves favorable performance against current state-of-the-art pansharpening methods. The code is available at https://github.com/coder-qicao/BiMPan.

Hyperspectral image super-resolution (HISR) is about fusing a low-resolution hyperspectral image (LR-HSI) and a high-resolution multispectral image (HR-MSI) to generate a high-resolution hyperspectral image (HR-HSI). Recently, convolutional neural network (CNN)-based techniques have been extensively investigated for HISR yielding competitive outcomes. However, existing CNN-based methods often require a huge amount of network parameters leading to a heavy computational burden, thus, limiting the generalization ability. In this article, we fully consider the characteristic of the HISR, proposing a general CNN fusion framework with high-resolution guidance, called GuidedNet. This framework consists of two branches, including 1) the high-resolution guidance branch (HGB) that can decompose the high-resolution guidance image into several scales and 2) the feature reconstruction branch (FRB) that takes the low-resolution image and the multiscaled high-resolution guidance images from the HGB to reconstruct the high-resolution fused image. GuidedNet can effectively predict the high-resolution residual details that are added to the upsampled HSI to simultaneously improve spatial quality and preserve spectral information. The proposed framework is implemented using recursive and progressive strategies, which can promote high performance with a significant network parameter reduction, even ensuring network stability by supervising several intermediate outputs. Additionally, the proposed approach is also suitable for other resolution enhancement tasks, such as remote sensing pansharpening and single-image super-resolution (SISR). Extensive experiments on simulated and real datasets demonstrate that the proposed framework generates state-of-the-art outcomes for several applications (i.e., HISR, pansharpening, and SISR). Finally, an ablation study and more discussions assessing, for example, the network generalization, the low computational cost, and the fewer network parameters, are provided to the readers. The code link is: https://github.com/Evangelion09/GuidedNet.

A Transformer has received a lot of attention in computer vision. Because of global self-attention, the computational complexity of Transformer is quadratic with the number of tokens, leading to limitations for practical applications. Hence, the computational complexity issue can be efficiently resolved by computing the self-attention in groups of smaller fixed-size windows. In this article, we propose a novel pyramid Shuffle-and-Reshuffle Transformer (PSRT) for the task of multispectral and hyperspectral image fusion (MHIF). Considering the strong correlation among different patches in remote sensing images and complementary information among patches with high similarity, we design Shuffle-and-Reshuffle (SaR) modules to consider the information interaction among global patches in an efficient manner. Besides, using pyramid structures based on window self-attention, the detail extraction is supported. Extensive experiments on four widely used benchmark datasets demonstrate the superiority of the proposed PSRT with a few parameters compared with several state-of-the-art approaches. The related code is available athttps://github.com/Deng-shangqi/PSRT.

Hyperspectral image super-resolution (HISR) aims to fuse a low-resolution image (LR-HSI) and a high-resolution multispectral image (HR-MSI), generating a high-resolution hyperspectral image (HR-HSI). Previous attempts to apply convolutional neural networks (CNNs) with spatial-variant adaptive filters for HISR tasks. Such filters overcome the spatial invariance and content-agnostic property of standard convolution. However, the current adaptive filters only consider pixellevel specificity, ignoring that each element of the features has unique close relationships with their neighbourhoods. To address the issue, we propose a context-aware element filter (CEF) operation, which generates adaptive filters for each element with sufficient perception of the specificity of each element to improve the representation capability. CEF can generate a single-channel filter to trade off the computational resource consumption for each element and is appropriate for HISR tasks with element-level dependencies. Specifically, we design a new network structure for HISR, which utilizes CEF to replace the standard convolution in the residual block. Extensive experiments demonstrate the superiority of the proposed CEF both visually and quantitatively compared with state-of-the-art (SOTA) methods.

Recently, convolutional neural networks (CNNs) have been introduced to pansharpening for enhancing fusion accuracy and overcoming the drawbacks of the conventional methods. However, most of methods based on CNN fail to distinguish the difference of multispectral bands, and only use a uniform set of convolutional kernels to extract features. In this paper, we design a progressive, band-separated convolutional network architecture for discriminatively learning the features and relation among spectral bands, aiming to address the problem mentioned before. More specifically, the proposed architecture mainly consists of three aspects. First, to accurately preserve the spectral peculiarities, we divide the multispectral input image in terms of its bands into several groups. Second, our original panchromatic and multispectral inputs are filtered by a high-pass operation to further yield more spatial details. Third, we use a spectral fusion module (SFM) for each group and associate them to progressively assemble the whole architecture. It is worth mentioning that the architecture could be integrated into any other competitive CNNs to improve the performance. Both visual and quantitative experiments have demonstrated that our proposed method outperforms recent state-of-the-art pansharpening techniques.