In remote sensing image scene classification (RSISC) tasks, downsampling is crucial for reducing computational complexity and cache demands, enhancing the model’s generalization capability of deep neural networks. Traditional methods, such as regular fixed lattice approaches (pooling in CNN and token merging in transformers), often flatten distinguishing texture features, impacting classification performance. To address this, we propose an adaptive token merging transformer (ATMformer) that preserves essential local features by estimating the importance score of each token. This allows significant tokens to be isolated during merging, mitigating the risk of feature blurring. Our experiments on three widely used RSISC datasets (NWPU-RESISC45, Aerial Image Dataset, and EuroSAT) demonstrate that ATMformer achieves state-of-the-art performance across all datasets. These findings underscore the importance of effective downsampling techniques in maintaining classification accuracy while efficiently processing large-scale data.

The process of fusing the rich spectral information of a low spatial resolution hyperspectral image (LR-HSI) with the spatial information of a high spatial resolution multispectral image (HR-MSI) to obtain an HSI with the spatial resolution of an MSI image is called hyperspectral image fusion (HIF). To reconstruct hyperspectral images at video frame rate, we propose a lightweight multi-level information network (MINet) for multispectral and hyperspectral image fusion. Specifically, we develop a novel lightweight feature fusion model, namely residual constraint block based on global variance fine-tuning (GVF-RCB), to complete the feature extraction and fusion of hyperspectral images. Further, we define a residual activity factor to judge the learning ability of the residual module, thereby verifying the effectiveness of GVF-RCB. In addition, we use cascade cross-level fusion to embed the different spectral bands of the upsampled LR-HSI in a progressive manner to compensate for lost spectral information at different levels and to maintain spatial high frequency information at all times. Experiments on different datasets show that our MINet outperforms the state-of-the-art methods in terms of objective metrics, in particular by requiring only 30% of the running time and 20% of the number of parameters.

Background: With the rapid development of high-throughput sequencing technology, the cost of whole genome sequencing drops rapidly, which leads to an exponential growth of genome data. How to efficiently compress the DNA data generated by large-scale genome projects has become an important factor restricting the further development of the DNA sequencing industry. Although the compression of DNA bases has achieved significant improvement in recent years, the compression of quality score is still challenging. Results: In this paper, by reinvestigating the inherent correlations between the quality score and the sequencing process, we propose a novel lossless quality score compressor based on adaptive coding order (ACO). The main objective of ACO is to traverse the quality score adaptively in the most correlative trajectory according to the sequencing process. By cooperating with the adaptive arithmetic coding and an improved in-context strategy, ACO achieves the state-of-the-art quality score compression performances with moderate complexity for the next-generation sequencing (NGS) data. Conclusions: The competence enables ACO to serve as a candidate tool for quality score compression, ACO has been employed by AVS(Audio Video coding Standard Workgroup of China) and is freely available at https://github.com/Yoniming/ACO.

Hyperspectral video acquisition requires a precise balance between spectral and temporal resolution, often achieved through compressive sampling using two-dimensional detectors and spectral reconstruction algorithms. However, the reliance on spatial light modulators for coding reduces optical efficiency, while complex recovery algorithms hinder real-time reconstruction. To address these challenges, we propose a digital-micromirror-device-based complementary dual-channel hyperspectral (DMD-CDH) video imaging system. This system employs a DMD for simultaneous light splitting and spatial encoding, enabling one channel to perform non-aliasing spectral sampling at lower frame rates while the other provides complementary high-rate sampling for panchromatic video. Featuring high optical throughput and efficient complementary sampling, the system ensures reliable hyperspectral video reconstruction and serves as a robust ground-based validation platform for remote sensing applications. Additionally, we introduce tailored optical error calibration and fixation techniques alongside a lightweight hyperspectral fusion network for reconstruction, achieving hyperspectral frame rates exceeding 30 fps. Compared to the existing models, this system simplifies the calibration process and provides a practical high-performance solution for real-time hyperspectral video imaging.

The basic principle of the patch-matching based style transfer is to substitute the patches of the content image feature maps by the closest patches from the style image feature maps. Since the finite features harvested from one single aesthetic style image are inadequate to represent the rich textures of the content natural image, existing techniques treat the full-channel style feature patches as simple signal tensors and create new style feature patches via signal-level fusion, which ignore the implicit diversities existed in style features and thus fail for generating better stylised results. In this paper, we propose a Retinex theory guided, channel-grouping based patch swap technique to solve the above challenges. Channel-grouping strategy groups the style feature maps into surface and texture channels, which prevents the winner-takes-all problem. Retinex theory based decomposition controls a more stable channel code rate generation. In addition, we provide complementary fusion and multi-scale generation strategy to prevent unexpected black area and overstylised results respectively. Experimental results demonstrate that the proposed method outperforms the existing techniques in providing more style-consistent textures while keeping the content fidelity.

Neural-network-based spectral fusion and reconstruction methods have achieved excellentperformance on simulated benchmarks, yet their deployment on real hardware systems re-mains challenging. The main obstacle is the mismatch between idealized data-driven assump-tions and the real sensing process: existing imaging models are often oversimplified, offlinecalibration cannot accommodate environment-induced parameter drift, and observation ex-traction becomes unreliable under model mismatch and sensor noise. To address these issues,this paper proposes an end-to-end physics-driven framework for real-system spectral imag-ing. We first define a Differentiable Spatially Varying Imaging Model (DSIM) that separatesfixed calibration parameters from learnable variables describing the dynamic system state.Based on this model, we then design an Observation Extraction Network (OENet) to esti-mate these dynamic parameters and recover reliable spectral observations from raw sensormeasurements, serving as a differentiable front-end for existing fusion backbones. We furtherintroduce a self-supervised online learning strategy based on 2D sensor-domain losses, en-abling post-deployment adaptation of both the extraction and fusion networks without spectrallabels. Through shared learnable parameters, the differentiable forward model, observationextraction, and online adaptation are unified into a physically consistent optimization pipelinefrom raw measurements to final reconstruction. Experimental results show that, when inte-grated with the proposed method, different fusion backbones achieve about a 3-dB PSNR gainon simulation datasets with real-world degradation, and experiments on a real DM-SCSI sys-tem further verify the effectiveness of the proposed framework.