Parameter-Efficient Fine-Tuning (PEFT) methods are crucial for adapting large pre-trained models. Among these, LoRA is considered a foundational approach. Building on this, the influential DoRA method enhances performance by decomposing weight updates into magnitude and direction. However, its underlying mechanism remains unclear, and it introduces significant computational overhead. In this work, we first identify that DoRA's success stems from its capacity to increase the singular value entropy of the weight update matrix, which promotes a more uniform update distribution akin to full fine-tuning. We then reformulate DoRA into a mathematically equivalent and more efficient matrix form, revealing it as a learnable weight conditioning method. Based on this insight, we propose a unified framework for designing advanced PEFT methods by exploring two orthogonal dimensions: the architectural placement and the transformation type of the conditioning matrix. Within this framework, we introduce two novel methods: (1) Pre-Diag, which applies a diagonal conditioning matrix before the LoRA update to efficiently calibrate the pre-trained weights, thereby enhancing performance while reducing training time; and (2) Skewed Orthogonal Rotation Adaptation (SORA), which employs a parameter-efficient orthogonal rotation to perform a more powerful, norm-preserving transformation of the feature space. Extensive experiments on natural language understanding and generation tasks demonstrate that our proposed methods achieve superior performance and efficiency compared to both LoRA and DoRA.

Despite the remarkable success of semantic token learning in NLP and vision domains, token-level representation mechanisms face fundamental challenges when extended to continuous time series analysis. We identify a core limitation lies in the intrinsic absence of semantically meaningful tokenization boundaries within time-series, which differs substantially from discrete text tokens and presents unique complexities compared to spatially coherent image patches. While existing works mechanically apply fixed-length partitioning, recent evidence from time series foundation models reveals performance ceilings in prediction tasks under such paradigms. This paper introduces a novel tokenization framework known as physics-aware tokenization (PATK), designed to implement adaptive time-frequency tokenization via distribution-sensitive sampling strategies. Key innovations include: 1) A Rate-of-Variation (RoV) distribution is meticulously structured to encompass multi-scale temporal dynamics in the time domain, alongside a Spectral Energy Intensity (SEI) distribution devised to reveal global seasonal patterns within the frequency domain; 2) A physics-aware hidden Markov modeling (PA-HMM) is then established to adaptively breaks down continuous time-series into distinct tokens with elastic lengths, responding to physics-aware probabilities sampled from RoV and SEI distributions. The proposed PATK allows steady integration with both conventional Transformers and advanced large-scale time series models (including LLM-transferred methods and pretrained time series foundation models). Simulations across various datasets demonstrate that PATK excels in classification and forecasting tasks, showing notable adaptability to model long-term dependencies, strengthening resilience against disturbances, and robustness to missing data events.

The absence of publicly available, large-scale, high-quality datasets for Synthetic Aperture Radar Automatic Target Recognition (SAR ATR) has significantly hindered the application of rapidly advancing deep learning techniques, which hold huge potential to unlock new capabilities in this field. This is primarily because collecting large volumes of diverse target samples from SAR images is prohibitively expensive, largely due to privacy concerns, the characteristics of microwave radar imagery perception, and the need for specialized expertise in data annotation. Throughout the history of SAR ATR research, there have been only a number of small datasets, mainly including targets like ships, airplanes, buildings, etc. There is only one vehicle dataset MSTAR collected in the 1990 s, which has been a valuable source for SAR ATR. To fill this gap, this paper introduces a large-scale, new dataset named ATRNet-STAR with 40 different vehicle categories collected under various realistic imaging conditions and scenes. It marks a substantial advancement in dataset scale and diversity, comprising over 190,000 well-annotated samples-$10\times$ larger than its predecessor, the famous MSTAR. Building such a large dataset is a challenging task, and the data collection scheme will be detailed. Secondly, we illustrate the value of ATRNet-STAR via extensively evaluating the performance of 15 representative methods with 7 different experimental settings on challenging classification and detection benchmarks derived from the dataset. Finally, based on our extensive experiments, we identify valuable insights for SAR ATR and discuss potential future research directions in this field. We hope that the scale, diversity, and benchmark of ATRNet-STAR can significantly facilitate the advancement of SAR ATR.

Hyperspectral image super-resolution (HSI-SR) is fundamentally more difficult than RGB image SR, since its ultrahigh spectral dimensionality. Existing supervised methods rely on labeled training data to obtain data prior, which incurs prohibitive collection costs and limits generalization. Unsupervised methods individually preset the band and kernel with handcrafted priors, whereas this decoupling modeling artificially creates a complexity-performance trade-off in the selected band number. To address these issues, we propose BKX-HMM, a unified statistical framework for blind HSI-SR, which uniformly models the band selection, kernel estimation, and HSI restoration through the state transition of a hidden Markov model (HMM). BKX-HMM redefines the trade-off as a distributional fitting problem: each Markov transition progressively learns optimal parameters of full-band distribution via limited spectral observations. Based on BKX-HMM, we propose BKSR, the first unsupervised blind HSI-SR method, which consists of three synergistic modules: Gibbs sampling-based band selection (GBS), test-time-training kernel estimation (TKE), and robust HSI restoration (RHR). These modules form a closed-loop optimization cycle: i) In GBS, the dynamic ergodicity of Gibbs sampling provides a global spectral view for kernel estimation and HSI restoration while maintaining local spectral computations; ii) In TKE, the GBS-sampled bands guide the kernel estimator update, achieving a learnable sampling-based mechanism, which refines kernel estimation to regularize RHR's diffusion trajectory; iii) In RHR, a spectral hyper-Laplacian prior is integrated into the reverse process of an off-the-shelf diffusion model, which achieves non-i.i.d. noise robust HSI restoration, feedback reweights band and kernel importance for subsequent GBS and TKE iterations. Extensive experiments on both synthetic and real HSI datasets demonstrate our BKSR's superiority over baseline methods across diverse scenarios (e.g., unknown Gaussian/motion kernel, non-i.i.d. noise) while maintaining comparable computational costs to the classic band selection methods.

Under the background of co-located MIMO radar, the existing reinforcement learning (RL)-based multi-target detection methods generally perform poorly on weak targets. In our previous work, we have proposed a beam optimization scheme with strong target limitation and gave a solution approach based on multi-rank beamformer, to achieve focusing more radar transmit power on weak targets. In this letter, we further propose a solution approach based on inner convex approximation, which can achieve a higher power gain due to its improved freedom. In addition, we also design an approach for choosing the focused angle cells of radar by fusing the statistical prior information from previous time. Summarizing the above improvements, we propose a RL-based multi-target detection method for MIMO radar assisted by strong target limitation. The experiments show that our method owns better performance on weak targets than its competitors while maintaining the excellent performance on strong targets.

In order to address the low radar-cross-section target detection challenge under high sea states and remove the dependence of existing methods on idealized prior knowledge, a two-stage closed-loop adaptive pulse-agile radar framework is proposed in this letter. First, the clutter and target parameters are estimated with the expectation-maximization scheme, followed by the processing of principal component analysis-based clutter suppression and generalized likelihood ratio test detection. Then, the multi-pulse polyphase-coded frequency modulate waveform is designed to minimize the complementary integrated sidelobe level in strong clutter regions with the iterative method. Numerical experiments are conducted to demonstrate the advantage of the proposed method in clutter suppression. To further confirm the robustness of our method, we also present the Monte Carlo experiments in challenging sea condition.

Target detection in aerial images captured by unmanned aerial vehicles has long been hampered by occlusion. Active Object Detection (AOD) aims to fundamentally address this issue from the active vision perspective, typically realized through the Deep Reinforcement Learning (DRL) paradigm. However, the active observation policy often suffers from low generalization ability, thus limiting its practical application. In this paper, we propose Divide-and-Conquer Sharpness-Aware Gradient Matching (DC-SAGM), a novel sharpness-based Domain Generalization (DG) method, to effectively enhance the generalization capacity of the agent’s policy. Specifically, we train the agent to learn the active observation policy using the conventional DRL approach. Sharpness-Aware Gradient Matching (SAGM) is employed during training, improving the model’s generalization performance by minimizing the sharpness metric of the loss landscape. Nevertheless, the imperfect state representation and classifier preference in the AOD problem lead to fierce gradient conflicts, deteriorating the effectiveness of SAGM. We address this incompatibility by using a divide-and-conquer strategy and exclude gradient conflicts via the majority-rule gradient surgery operation. Extensive experimental results on the UEVAVD dataset validate DC-SAGM’s superiority in helping the agent’s policy achieve better generalization compared to extensive policy learning approaches.

Infrared and visible image fusion methods have shown promising results, yet existing approaches either compromise downstream detection performance through independent fusion processes or sacrifice computational efficiency and flexibility by requiring joint training of fusion and detection models. To address these challenges, we propose a detection-driven image fusion network based on diffusion models (termed as DDIF), which optimizes the fused images specifically for object detection tasks. Our method features the following three aspects: 1) we reformulate the image fusion process as an inverse problem solved by a non-differentiable optimization process wherein the fused result preserves the source modality information while conforming to the image prior provided by the diffusion model; 2) we design a Response Guide Learning Module (RGLM) to learn response maps, which determine the contribution of each modality in the fusion process according to the downstream detection task; 3) we establish explicit gradient relationships to ensure compatibility between RGLM training and the non-differentiable optimization process, enabling end-to-end training. Notably, a moderate coupling mechanism is formed in our framework as the subsequent detection model is pre-trained and frozen, enabling flexible integration with various advanced detection networks while maintaining computational efficiency. Extensive experiments indicate that our method achieves superior detection performance compared to SOTA approaches and produces high-quality image fusion results.

Unmanned aerial vehicles (UAV)-based object detection with visible (RGB) and infrared (IR) images facilitates robust around-the-clock detection, driven by advancements in deep learning techniques and the availability of high-quality dataset. However, the existing dataset struggles to fully capture real-world complexity for limited imaging conditions. To this end, we introduce a high-diversity dataset ATR-UMOD covering varying scenarios, spanning altitudes from 80m to 300m, angles from 0° to 75°, and all-day, all-year time variations in rich weather and illumination conditions. Moreover, each RGB-IR image pair is annotated with 6 condition attributes, offering valuable high-level contextual information. To meet the challenge raised by such diverse conditions, we propose a novel prompt-guided condition-aware dynamic fusion (PCDF) to adaptively reassign multimodal contributions by leveraging annotated condition cues. By encoding imaging conditions as text prompts, PCDF effectively models the relationship between conditions and multimodal contributions through a task-specific soft-gating transformation. A prompt-guided condition-decoupling module further ensures the availability in practice without condition annotations. Experiments on ATR-UMOD dataset reveal the effectiveness of PCDF.

Accurate radio environment maps (REMs) can enhance the performance of wireless networks and optimize spectrum utilization efficiency. However, in rugged terrain environments, radio propagation is significantly affected by terrain variations, resulting in spatial heterogeneity in received signal strength (RSS) and impairing the accuracy of REM construction. To address these challenges, a Gaussian Process Regression method incorporating terrain (GPRT) is proposed to exploit both spatial and terrain correlation properties. In GPRT, a specialized kernel function is designed to integrate digital elevation data into the Gaussian process framework, capturing anisotropic spatial correlation and terrain effects. In addition, an Adaptive Moment Estimation (Adam) optimization algorithm is utilized for efficient hyperparameter tuning, enhancing convergence speed and parameter accuracy. Simulations with varying numbers of emitters and field experiments in real-world terrain demonstrate the superiority and effectiveness of the proposed GPRT over competing methods in terms of robustness and accuracy. Specifically, GPRT outperformed the best comparative approaches by 20% to 33% in simulations and by up to 20% in the field experiment.

This paper addresses the issue of distributed fuzzy fault-tolerant formation control (FTFC) for networked unmanned aerial vehicles (UAVs), considering the need for resilient prescribed performance under saturation constraints. First, an online self-organizing interval type-2 fuzzy neural network (OSITFNN) is developed to handle the compensation of unknown terms, which can adaptively update the network structure during the control process to obtain better dynamic mapping. Subsequently, a resilient prescribed performance function (RPPF) is designed for automatically scaling performance boundaries based on the saturation level of the control input, reducing the conflict between actuator saturation and performance constraints. This leads to the proposal of a distributed resilient fractional-order sliding mode FTFC (DRFTFC) scheme, ensuring stable formation control. The uncertain fuzzy network weights are compensated through a series of adaptive laws. Simulation results are provided to validate the effectiveness and robustness of the proposed control scheme.

Incorporating heterogeneous representations from different architectures has facilitated various vision tasks, e.g., some hybrid networks combine transformers and convolutions. However, complementarity between such heterogeneous architectures has not been well exploited in self-supervised learning. Thus, we propose Heterogeneous Self-Supervised Learning (HSSL), which enforces a base model to learn from an auxiliary head whose architecture is heterogeneous from the base model. In this process, HSSL endows the base model with new characteristics in a representation learning way without structural changes. To comprehensively understand the HSSL, we conduct experiments on various heterogeneous pairs containing a base model and an auxiliary head. We discover that the representation quality of the base model moves up as their architecture discrepancy grows. This observation motivates us to propose a search strategy that quickly determines the most suitable auxiliary head for a specific base model to learn and several simple but effective methods to enlarge the model discrepancy. The HSSL is compatible with various self-supervised methods, achieving superior performances on various downstream tasks, including image classification, semantic segmentation, instance segmentation, and object detection.

While recent debiasing methods for Scene Graph Generation (SGG) have shown impressive performance, these efforts often attribute model bias solely to the long-tail distribution of relationships, overlooking the more profound causes stemming from skewed object and object pair distributions. In this paper, we employ causal inference techniques to model the causality among these observed skewed distributions. Our insight lies in the ability of causal inference to capture the unobservable causal effects between complex distributions, which is crucial for tracing the roots of model bias. Specifically, we introduce the Mediator-based Causal Chain Model (MCCM), which, in addition to modeling causality among objects, object pairs, and relationships, incorporates mediator variables, i.e., cooccurrence distribution, for complementing the causality. Following this, we propose the Causal Adjustment Module (CAModule) to estimate the modeled causal structure, using variables from MCCM as inputs to produce a set of adjustment factors aimed at correcting biased model predictions. Moreover, our method enables the composition of zero-shot relationships, thereby enhancing the model's ability to recognize such relationships. Experiments conducted across various SGG backbones and popular benchmarks demonstrate that CAModule achieves state-of-the-art mean recall rates, with significant improvements also observed on the challenging zero-shot recall rate metric.

Existing two-stage Scene Graph Generation (SGG) frameworks typically incorporate a detector to extract relationship features and a classifier to categorize these relationships; therefore, the training paradigm follows a causal chain structure, where the detector's inputs determine the classifier's inputs, which in turn influence the final predictions. However, such a causal chain structure can yield spurious correlations between the detector's inputs and the final predictions, i.e., the prediction of a certain relationship may be influenced by other relationships. This influence can induce at least two observable biases: tail relationships are predicted as head ones, and foreground relationships are predicted as background ones; notably, the latter bias is seldom discussed in the literature. To address this issue, we propose reconstructing the causal chain structure into a reverse causal structure, wherein the classifier's inputs are treated as the confounder, and both the detector's inputs and the final predictions are viewed as causal variables. Specifically, we term the reconstructed causal paradigm as the Reverse causal Framework for SGG (RcSGG). RcSGG initially employs the proposed Active Reverse Estimation (ARE) to intervene on the confounder to estimate the reverse causality, i.e., the causality from final predictions to the classifier's inputs. Then, the Maximum Information Sampling (MIS) is suggested to enhance the reverse causality estimation further by considering the relationship information. Theoretically, RcSGG can mitigate the spurious correlations inherent in the SGG framework, subsequently eliminating the induced biases. Comprehensive experiments on popular benchmarks and diverse SGG frameworks show the state-of-the-art mean recall rate.

The fundamental challenge in SAR target detection lies in developing discriminative, efficient, and robust representations of target characteristics within intricate non-cooperative environments. However, accurate target detection is impeded by factors including the sparse distribution and discrete features of the targets, as well as complex background interference. In this study, we propose a Gamma Diffusion Model Network with MambaSAR module (MaDiNet) for SAR target detection. Specifically, MaDiNet leverages the Gamma distribution to model the statistical characteristics of SAR images, and conceptulizes SAR target detection as the task of generating target bounding boxes in the image space. Furthermore, we design a MambaSAR module to capture intricate spatial structural information of targets and enhance the capability of the model to differentiate between targets and complex backgrounds. The experimental results on multi-class target detection datasets have all achieved SOTA, with a particularly notable improvement of 6.7% in mAP50 on the ODSOG-1.0 dataset, proving the effectiveness of the proposed network. Code is available at https://github.com/JoyeZLearning/MaDiNet.

In passive ground target detection using Unmanned Aerial Vehicles (UAVs), some detrimental factors like occlusion significantly impact target detection performance. Active Object Detection offers an effective way to address it, which usually uses Deep Reinforcement Learning (DRL) to plan UAV’s viewpoint for favorable observations. However, existing DRL-based AOD methods often suffer from low sample efficiency and poor generalization due to inadequate state representation learned by the policy network. Inspired by human scene understanding where their spatial representation of the scene remains consistent despite different observation orders, we design a self-supervised state representation learning method based on Observations Temporal Permutation (OTP) to improve the state representation of the agent’s policy network. We require the policy network to output consistent action value estimates for observation sequences with the same content but different temporal orders. Besides, we use the state representation to predict the target orientation variations in the observation sequence, which further regularizes and facilitates the state representation learning process. Finally, we design multiple experiments based on the UEVAVD dataset to compare the proposed method with existing self-supervised state representation learning methods for the AOD task. The experimental results demonstrate that the OTP method can help the agent’s policy network learn a better state representation, thus achieving higher policy learning sample efficiency and stronger policy generalization.

Dispersive processing of borehole acoustic logging data is essential for accurate evaluation of subsurface formation velocities, which directly impact reservoir characterization and geophysical interpretation. Traditional coherence-based techniques fail to address the dispersive nature of guided waves, leading to underestimated velocities, while model-based inversion requires strict parameter assumptions that limit applicability in complex formations. To overcome these challenges, we propose a physics-informed multi-task deep learning method that integrates an analytical exponential function, Gaussian noise modeling, and a hybrid CNN–BiLSTM architecture. Large-scale synthetic datasets are generated separately for fast and and slow formation scenarios, enabling the network to capture both local dispersion curve details and long-range sequential dependencies. A composite physics-informed loss function jointly constrains parameter accuracy and dispersion curve consistency, ensuring physically meaningful predictions. Numerical experiments on simulated dipole waveforms demonstrate that the proposed method can accurately reconstruct clean dispersion curves and invert reliable formation shear velocities. Furthermore, field tests on LWD quadrupole wave data validate its robustness and generalization capability, confirming its practical potential for borehole acoustic logging applications.

Registration of optical and synthetic aperture radar (SAR) remote sensing images serves as a critical foundation for image fusion and visual navigation tasks. This task is particularly challenging because of their modal discrepancy, primarily manifested as severe nonlinear radiometric differences (NRD), geometric distortions, and noise variations. Under large geometric transformations, existing classical template-based and sparse keypoint-based strategies struggle to achieve reliable registration results for optical-SAR image pairs. To address these limitations, we propose GDROS, a geometry-guided dense registration framework leveraging global cross-modal image interactions. First, we extract cross-modal deep features from optical and SAR images through a CNN-Transformer hybrid feature extraction module, upon which a multi-scale 4D correlation volume is constructed and iteratively refined to establish pixel-wise dense correspondences. Subsequently, we implement a least squares regression (LSR) module to geometrically constrain the predicted dense optical flow field. Such geometry guidance mitigates prediction divergence by directly imposing an estimated affine transformation on the final flow predictions. Extensive experiments have been conducted on three representative datasets WHU-Opt-SAR dataset, OS dataset, and UBCv2 dataset with different spatial resolutions, demonstrating robust performance of our proposed method across different imaging resolutions. Qualitative and quantitative results show that GDROS significantly outperforms current state-of-the-art methods in all metrics. Our source code will be released at: https://github.com/Zi-Xuan-Sun/GDROS.

Recent research in remote sensing object detection (RSOD) has significantly advanced the development of vision foundation models. However, deploying these models on resource-constrained edge devices is challenging due to their high computational demands. Binarized detectors utilize binary neural networks (BNNs) to achieve extreme compression by quantizing weights and activations to +1 or −1, which have been extensively studied for generic object detection tasks. In remote sensing images, the objects of interest typically exhibit weak responses, and the images often contain numerous unique local areas. Feature binarization in these images can lead to substantial loss of object contrast and scale prior information, which exacerbates performance issues, particularly for small objects, resulting in significant performance degradation. To address these challenges, we propose a novel binarized detector for RSOD named the heterogeneous binary pixel difference network (HBiPiDiNet). Initially, we developed a binary pixel difference convolution (BiPDC) that integrates local binary patterns (LBPs) to capture local contrast information with traditional binary convolution, thereby enhancing the representation of small objects. Subsequently, we constructed heterogeneous kernel fusion convolution blocks (HKFCB) based on BiPDC and standard binary convolution. The HKFCB comprises multiple BiPDCs at different scales, effectively representing BiPDC under multiscale LBP and multiscale binary convolutions. Extensive experiments demonstrate that our proposed method significantly enhances the performance of state-of-the-art binary detection methods across three remote sensing datasets: AI-TOD, VisDrone2019, and DIOR. We have released our code and models athttps://github.com/yuhua666/HBiPiDiNet/tree/main.

Mathematically, the autofocusing and imaging model for sparse aperture inverse synthetic aperture radar (ISAR) has an infinite number of solutions, even with the addition of some sparsity constraints. What relationship exists between these infinite solutions should be answered. In addition, some existing approaches suffer from troublesome manual fine-tuning parameters, or high algorithmic complexity, or low reconstruction accuracy. To deal with the above problems, a sparse aperture ISAR autofocusing and imaging method combining phase estimation and log-sum minimization is proposed in this paper, which has high estimation accuracy and computational efficiency, and avoids complicated manual parameter adjustment. Focusing on the model with a phase diagonal matrix and log-sum function, we reveal why there are countless solutions and what relationships exist between them. These features are shared by other models including different regularization functions. Without sparsity prior, the regularization parameter of the algorithm is changed automatically at each iteration accroding to the ratio of the selected element to the maximum absolute value element in an auxiliary matrix, which controls that only a few elements participate in the computation at each iteration. Coupled with the situation that computationally heavy matrix inversion operations are not required, the calculation speed of the approach is greatly boosted. To show the stability of the algorithm, its convergence is proven without any assumptions. Experiments based on both simulation and measurement data demonstrate that compared with the recently proposed algorithms by others, the proposed algorithm can achieve well-focused ISAR images in less than one second and is very efficient to implement.

Most state-of-the-art object detection methods suffer from poor generalization due to the domain shift between training and testing datasets. To resolve this challenge, unsupervised cross domain object detection is proposed to learn an object detector for an unlabeled target domain by transferring knowledge from an annotated source domain. Promising results have been achieved via Mean Teacher, however, pseudo labeling which is the bottleneck of mutual learning remains to be further explored. In this study, we find that confidence misalignment of the predictions, including category-level overconfidence, instance-level task confidence inconsistency, and image-level confidence misfocusing, leading to the injection of noisy pseudo labels in the training process, will bring suboptimal performance. Considering the above issue, we present a novel general framework termed Multi-Granularity Confidence Alignment Mean Teacher (MGCAMT) for unsupervised cross domain object detection, which alleviates confidence misalignment across category-, instance-, and image-levels simultaneously to refine pseudo labeling for better teacher-student learning. Specifically, to align confidence with accuracy at category level, we propose Classification Confidence Alignment (CCA) to model category uncertainty based on Evidential Deep Learning (EDL) and filter out the category incorrect labels via an uncertainty-aware selection strategy. Furthermore, we design Task Confidence Alignment (TCA) to mitigate the instance-level misalignment between classification and localization by enabling each classification feature to adaptively identify the optimal feature for regression. Finally, we develop imagery Focusing Confidence Alignment (FCA) adopting another way of pseudo label learning, i.e., we use the original outputs from the Mean Teacher network for supervised learning without label assignment to achieve a balanced perception of the image's spatial layout. When these three procedures are integrated into a single framework, they mutually benefit to improve the final performance from a cooperative learning perspective. Extensive experiments across multiple scenarios demonstrate that our method outperforms large foundational models, and surpasses other state-of-the-art approaches by a large margin.

Despite the remarkable progress in synthetic aperture radar automatic target recognition (SAR ATR), recent efforts have concentrated on detecting and classifying a specific category, e.g., vehicles, ships, airplanes, or buildings. One of the fundamental limitations of the top-performing SAR ATR methods is that the learning paradigm is supervised, task-specific, limited-category, closed-world learning, which depends on massive amounts of accurately annotated samples that are expensively labeled by expert SAR analysts and have limited generalization capability and scalability. In this work, we make the first attempt towards building a foundation model for SAR ATR, termed SARATR-X. SARATR-X learns generalizable representations via self-supervised learning (SSL) and provides a cornerstone for label-efficient model adaptation to generic SAR target detection and classification tasks. Specifically, SARATR-X is trained on 0.18 M unlabelled SAR target samples, which are curated by combining contemporary benchmarks and constitute the largest publicly available dataset till now. Considering the characteristics of SAR images, a backbone tailored for SAR ATR is carefully designed, and a two-step SSL method endowed with multi-scale gradient features was applied to ensure the feature diversity and model scalability of SARATR-X. The capabilities of SARATR-X are evaluated on classification under few-shot and robustness settings and detection across various categories and scenes, and impressive performance is achieved, often competitive with or even superior to prior fully supervised, semi-supervised, or self-supervised algorithms. Our SARATR-X and the curated dataset are released at https://github.com/waterdisappear/SARATR-X to foster research into foundation models for SAR image interpretation.

Pan-sharpening is a task that aims to super-resolve the low-resolution multispectral (LRMS) image with the guidance of a corresponding high-resolution panchromatic (PAN) image. The key challenge in pan-sharpening is to accurately modeling the relationship between the MS and PAN images. While supervised deep learning methods are commonly employed to address this task, the unavailability of ground-truth severely limits their effectiveness. In this paper, we propose a mutually guided detail restoration method for unsupervised pan-sharpening. Specifically, we treat pan-sharpening as a blind image deblurring task, in which the blur kernel can be estimated by a CNN. Constrained by the blur kernel, the pan-sharpened image retains spectral information consistent with the LRMS image. Once the pan-sharpened image is obtained, the PAN image is blurred using a pre-defined blur operator. The pan-sharpened image, in turn, is used to guide the detail restoration of the blurred PAN image. By leveraging the mutual guidance between MS and PAN images, the pan-sharpening network can implicitly learn the spatial relationship between the two modalities. Extensive experiments show that the proposed method significantly outperforms existing unsupervised pan-sharpening methods.

Deep learning-based methods have achieved significant successes on solving the blind super-resolution (BSR) problem. However, most of them request supervised pretraining on labelled datasets. This paper proposes an unsupervised kernel estimation model, named dynamic kernel prior (DKP), to realize an unsupervised and pretraining-free learning-based algorithm for solving the BSR problem. DKP can adaptively learn dynamic kernel priors to realize real-time kernel estimation, and thereby enables superior HR image restoration performances. This is achieved by a Markov chain Monte Carlo sampling process on random kernel distributions. The learned kernel prior is then assigned to optimize a blur kernel estimation network, which entails a network-based Langevin dynamic optimization strategy. These two techniques ensure the accuracy of the kernel estimation. DKP can be easily used to replace the kernel estimation models in the existing methods, such as Double-DIP and FKP-DIP, or be added to the off-the-shelf image restoration model, such as diffusion model. In this paper, we incorporate our DKP model with DIP and diffusion model, referring to DIP-DKP and Diff-DKP, for validations. Extensive simulations on Gaussian and motion kernel scenarios demonstrate that the proposed DKP model can significantly improve the kernel estimation with comparable runtime and memory usage, leading to state-of-the-art BSR results. The code is available at https://github.com/XYLGroup/DKP.

Mix-precision computation is crucial for artificial intelligence and scientific computing applications. However, as novel chips with innovative architectures emerge, harnessing their computational capabilities presents significant challenges. While existing algorithms for the HPL-MxP LU factorization excel on homogeneous systems, they often encounter difficulties on specialized heterogeneous architectures. This deficiency arises from inadequate optimization for computation, memory access, and communication, hindering effective mixed-precision acceleration. This work introduces an algorithm-hardware co-optimization approach for LU factorization on specialized NPUs and CPUs, leveraging their unique architectures. A novel multi-iteration fusion method for general matrix multiplication is proposed, strategically designed to maximize on-chip L1 buffer utilization, effectively overcoming the notorious “memory wall”. Additionally, a multi-stage, multi-level heterogeneous pipeline for LU factorization in an accelerator-CPU cloud environment is presented, where compute-intensive matrix multiplications are offloaded to NPUs while CPUs handle the remaining tasks. The co-optimization approach fosters deep collaboration between CPUs and accelerators, thereby unlocking enhanced performance.

Registration of optical and synthetic aperture radar (SAR) image pairs is a fundamental task in various remote sensing applications, including image fusion, target localization, and object detection. Unlike homogeneous image pairs, optical and SAR image pairs exhibit a significant modality gap, making it exceptionally challenging to extract consistent and reliable features. Particularly for optical and SAR image pairs with substantial geometric differences, few methods can achieve high-precision registration. To address this challenging task, we introduce a novel registration framework, called OS3Flow, leveraging on the implicit symmetry between heterogeneous image pairs to extract high-quality semi-dense flow estimations. We start by training the network in a multi-task manner using a standard flow regression loss as well as a symmetry loss with reverse input order. A confidence mask thus can be generated to measure the similarity between predictions at inference time. We then perform a linear regression upon selected flows with high confidence to estimate the parameters of underlying affine transformation. Under large transformations, our proposed method achieves an average registration error of less than 3 pixels on the public OS dataset and WHU-OPT-SAR dataset, demonstrating superior accuracy and robustness compared to state-of-the-art methods.

Under the background of slow small target detection on sea surface, it is a serious problem that the detection performance of existing feature detection methods decreases when the number of coherent pulses is less. In this letter, we first model the slow small target detection on sea surface as a contextual bandit problem. On this basis, we propose an autonomous feature detection method by modifying the classical feature detection process. The method can autonomously choose the detectors with better performance under current sea scene from the constructed 18 feature detectors, and obtain fine detection performance by fusing their detection results. The performance superiority and the real-time capability of proposed method are verified by the experiments on 7 CSIR datasets.

Sparse aperture ISAR imaging for maneuvering targets is a relatively difficult task due to the complex form of observation model. In this letter, an approximated observation model based on CZT and NCS is proposed to accelerate the implementation of forward and backward operators. A structured sparse prior is introduced to establish a statistical framework for SA-ISAR imaging and VB-GAMP is utilized to implement a fast inference. A rotation parameters estimation based on image quality optimization is further plugged in the reconstruction procedure to achieve joint imaging and motion compensation. Experiments on simulated and measured data validate the effectiveness and efficiency of the proposed method.

Aircraft target detection in SAR images is a challenging task due to the discrete scattering points and severe background clutter interference. Currently, methods with convolution-based or transformer-based paradigms cannot adequately address these issues. In this letter, we explore diffusion models for SAR image aircraft target detection for the first time and propose a novel Diffusion-based aircraft target Detection network for SAR images (DiffDet4SAR). Specifically, the proposed DiffDet4SAR yields two main advantages for SAR aircraft target detection: 1) DiffDet4SAR maps the SAR aircraft target detection task to a denoising diffusion process of bounding boxes without heuristic anchor size selection, effectively enabling large variations in aircraft sizes to be accommodated; and 2) the dedicatedly designed Scattering Feature Enhancement (SFE) module further reduces the clutter intensity and enhances the target saliency during inference. Extensive experimental results on the SAR-AIRcraft-1.0 dataset show that the proposed DiffDet4SAR achieves 88.4% mAP50, outperforming the state-of-the-art methods by 6%. Code is availabel at https://github.com/JoyeZLearning/DiffDet4SAR.

A sparsity-based adaptive beamforming (ABF) method is introduced to effectively process coherent signals with polarized sensor arrays (PSA). This method exploits the spatial sparsity of observed signals by transforming it into row-sparsity within a waveform-polarization composite matrix through data reorganization. This row-sparsity is subsequently cast as an$\ell _{2,1}$norm minimization problem, characterized by a gridless and compact mathematical expression with a Hermitian Toeplitz matrix. Then, a matrix factorization-based gradient descent (GD) algorithm is introduced to effectively resolve this optimization problem. The experimental evaluations demonstrate that the GD algorithm significantly outperforms the MOSEK solver in terms of computational efficiency. Further comparative analysis demonstrates that the proposed method outperforms the existing techniques, especially in contexts of low signal-to-noise ratio (SNR), with a moderate increase in computational runtime.

This paper focuses on generating Inverse Synthetic Aperture Radar (ISAR) images from optical images, in particular, for orbit space targets. ISAR images are widely applied in space target observation and classification tasks, whereas, limited to the expensive cost of ISAR sample collection, training deep learning-based ISAR image classifiers with insufficient samples and generating ISAR samples from emulation optical images via image translation techniques have attracted increasing attention. Image translation has highlighted significant success and popularity in computer vision, remote sensing and data generation societies. However, most of the existing methods are implemented under the discipline of extracting the explicit pixel-level features and do not perform effectively while entailing translation to domains with specific implicit features, such as ISAR image does. We propose a meta-learning based domain prior to implicit feature modelling and apply it to CycleGAN and UNIT models to realize effective translations between the ISAR and optical domains. Two representative implicit features, ISAR scattering distribution feature from the physical domain and the classification identifying feature from the task domain, are elaborately formulated with explicit modelling in statistic form. A meta-learning based training scheme is introduced to leverage the mutual knowledge of domain priors across different samples, and thus allows few-shot learning capacity with dramatically reduced training samples. Extensive simulations validate that the obtained ISAR images have better visible-authenticity and training-effectiveness than the existing image translation approaches on various synthetic datasets. Source codes are available at.

This article focuses on the multiazimuth angle interpolation task of inverse synthetic aperture radar (ISAR) images for aircraft targets and complements incomplete ISAR image datasets. ISAR image automatic target recognition (ATR) has been widely applied in remote sensing and many fields. However, the imaging process is more challenging when compared to capturing optical and SAR image data, which reduces the accuracy and generalization performance of the ATR system. Therefore, in this article, we leverage existing limited ISAR data to achieve autonomous data expansion. This approach helps mitigate the impact of low sample quantity and unbalanced distribution, ultimately improving the accuracy of the ATR system for target recognition. Most existing methods use generative networks for ISAR image expansion, but few focus on generating ISAR images with specific azimuth angles. This article proposes a novel two-stage coarse-to-fine framework for ISAR object view interpolation (C2FIPNet) that combines flow estimation and GAN to interpolate ISAR images with intermediate azimuth angles using a set of ISAR image pairs. Flow estimation is employed for coarse-grained generation, determining the position and intensity of strong scattering points in the ISAR image. The GAN, on the other hand, is used for fine-grained completion to correct image distortion caused by flow estimation and enhance image details. In addition, a suitable loss function is designed, incorporating both global and local features, allowing for priority generation in the region of strong scattering points. In conclusion, extensive simulation and comparative experiments have demonstrated that the interpolated ISAR images generated by the proposed C2FIPNet exhibit greater pixel-level authenticity.

For inverse synthetic aperture radar (ISAR) imaging under sparse aperture (SA) conditions, the rotation motion compensation is seldom considered. However, with the improvement of resolution, the migration through resolution cell (MTRC) cannot be ignored. Traditional methods for rotation motion compensation generally fail in SA cases. This article proposes a method to jointly implement sparse imaging and compensation of the MTRC in a structured sparse Bayesian learning (SBL) framework. Due to the coupling of fast time and slow time, the observation model is established in a vectorized form. To reduce the computational complexity, approximated inference methods are utilized to achieve fast inference for the posteriors. Maximum contrast (MC) criterion is adopted to estimate the rotation parameters. The approximated implementation for the forward operator and backward operator is discussed to further accelerate the algorithm. Experimental results based on simulated and measured data validate the effectiveness and efficiency of the proposed methods.

Existing Cross-Domain Few-Shot Learning (CDFSL) methods require access to source domain data to train a model in the pre-training phase. However, due to increasing concerns about data privacy and the desire to reduce data transmission and training costs, it is necessary to develop a CDFSL solution without accessing source data. For this reason, this paper explores a Source-Free CDFSL (SF-CDFSL) problem, in which CDFSL is addressed through the use of existing pretrained models instead of training a model with source data, avoiding accessing source data. However, due to the lack of source data, we face two key challenges: effectively tackling CDFSL with limited labeled target samples, and the impossibility of addressing domain disparities by aligning source and target domain distributions. This paper proposes an Enhanced Information Maximization with Distance-Aware Contrastive Learning (IM-DCL) method to address these challenges. Firstly, we introduce the transductive mechanism for learning the query set. Secondly, information maximization (IM) is explored to map target samples into both individual certainty and global diversity predictions, helping the source model better fit the target data distribution. However, IM fails to learn the decision boundary of the target task. This motivates us to introduce a novel approach called Distance-Aware Contrastive Learning (DCL), in which we consider the entire feature set as both positive and negative sets, akin to Schrödinger's concept of a dual state. Instead of a rigid separation between positive and negative sets, we employ a weighted distance calculation among features to establish a soft classification of the positive and negative sets for the entire feature set. We explore three types of negative weights to enhance the performance of CDFSL. Furthermore, we address issues related to IM by incorporating contrastive constraints between object features and their corresponding positive and negative sets. Evaluations of the 4 datasets in the BSCD-FSL benchmark indicate that the proposed IM-DCL, without accessing the source domain, demonstrates superiority over existing methods, especially in the distant domain task. Additionally, the ablation study and performance analysis confirmed the ability of IM-DCL to handle SF-CDFSL. The code will be made public at https://github.com/xuhuali-mxj/IM-DCL.

Neuro-Rule Networks (NRNs) emerge as a promising neuro-symbolic method, enjoyed by the ability to equate fully-connected neural networks with logic rules. To support learning logic rules consisting of boolean variables, converting input features into binary representations is required. Different from discrete features that could be directly transformed by one-hot encodings, continuous features need to be binarized based on some numerical intervals. Existing studies usually select the bound values of intervals based on empirical strategies (e.g., equal-width interval). However, it is not optimal since the bounds are fixed and cannot be optimized to accommodate the ultimate training target. In this paper, we propose AutoInt, an approach that automatically binarizes continuous features and enables the intervals to be optimized with NRNs in an end-to-end fashion. Specifically, AutoInt automatically selects an interval for a given continuous feature in a soft manner to enable a differentiable learning procedure of interval-related parameters. Moreover, it introduces an additional soft K-means clustering loss to make the interval centres approach the original feature value distribution, thus reducing the risk of overfitting intervals. We conduct comprehensive experiments on public datasets and demonstrate the effectiveness of AutoInt in boosting the performance of NRNs.

The inverse synthetic aperture radar (ISAR) images are often afflicted by boundary blurring, discontinuity, sidelobe effects of strong scattering points, a large dynamic range of gray values, and azimuth defocus, which pose significant challenges to image segmentation. This paper proposes a novel semantic segmentation method for ISAR images of space targets. The method is based on contrastive learning (CL) and Non-Local Unet (NL-Unet). First, the method roughly segments the target contour using binary semantic tags to remove sidelobe interference and image noise. Then, the Non-local self-attentive mechanism with a global perceptual field is used to exploit the structural symmetry of the ISAR image. Finally, to improve the segmentation ability of target small parts, the method adopts a training method based on CL to overcome the relatively weak problem of the supervised learning model. The proposed method outperforms existing methods on the simulated ISAR dataset. Moreover, it is practical and can be directly generalized to the real-measured dataset without retraining.

In recent years, deep learning has been widely used in SAR ATR and achieved excellent performance on the MSTAR dataset. However, due to constrained imaging conditions, MSTAR has data biases such as background correlation,i.e., background clutter properties have a spurious correlation with target classes. Deep learning can overfit clutter to reduce training errors. Therefore, the degree of overfitting for clutter reflects the non-causality of deep learning in SAR ATR. Existing methods only qualitatively analyze this phenomenon. In this paper, we quantify the contributions of different regions to target recognition based on the Shapley value. The Shapley value of clutter measures the degree of overfitting. Moreover, we explain how data bias and model bias contribute to non-causality. Concisely, data bias leads to comparable signal-to-clutter ratios and clutter textures in training and test sets. And various model structures have different degrees of overfitting for these biases. The experimental results of various models under standard operating conditions on the MSTAR dataset support our conclusions. Our code is available at https://github.com/waterdisappear/Data-Bias-in-MSTAR.

A multistatic ISAR system can observe a target from multiple observation angles. Compared with the monostatic ISAR system, the multistatic ISAR system can obtain more spatial sampling data, which provides the ability for high-resolution ISAR imaging. In some cases, the locations of radars are close. There are overlaps among observing angles, which brings little cross-range resolution improvement. However, such scenes are less considered in previous work. In the scene with overlaps, the image obtained by each radar may be similar due to the similar observation angles. In this letter, a novel multistatic ISAR imaging model is proposed by applying the similarity prior as a constraint. And an effecient image reconstruction algorithm is derived based on the orthogonality of observation matrix. Compared with existing CS based methods, the proposed method can be directly applied on multistatic ISAR echoes without pre-processing of rearranging, which is more convenient in practical applications. Experiment results of simulated and measured data show that the proposed method achieves better performance especially under low signal-to-noise (SNR) conditions.

A sparse reconstruction approach for adaptive beamforming (ABF) with polarization-sensitive arrays (PSA) is introduced in this letter. It first represents the spatial sparsity of incoming signals as the row sparsity of a power-scaled polarization matrix, which arises from the matrization of the redundancy-reduced covariance vector. Then the row sparsity issue is relaxed to an$\ell _{2,1}$norm minimization form and solved in a gridless way via a compact formulation, where a dimension reduction method is introduced to reduce the problem size. Compared to existing techniques, the proposed method processes the polarization information holistically and derives each signal parameter in the continuous domain. Simulation results substantiate the advantages of the proposed method over competing methods.

Precession is a typical form of micro-motion that can bring about complex and time-varying Doppler modulation. The range instantaneous Doppler (RID) method, which uses time-frequency analysis instead of the Fourier transform to describe the time-varying Doppler, is typically used to obtain the high-resolution inverse synthetic aperture radar (ISAR) image of a precession target. However, the observation time of a specific target is often non-uniform due to various interference and channel switching among multi-channel radars, which will lead to a sparse aperture. Sparse aperture can cause sidelobe interference in the ISAR image obtained by the RID method, making it difficult to focus well. To solve the problem whereby the RID method fails to image a precession target with sparse aperture, this paper proposes a new method based on the joint constraints of low-rank and sparsity of tensor, and uses the alternating direction method of multipliers to solve the problem. The low-rank can constrain the correlation among consecutive ISAR images, and sparsity can remove the impact of sparse aperture. This effectively eliminates the micro-Doppler interference and sidelobe interference in ISAR image, and enables the reconstruction of precession target in a sequence of ISAR images with sparse aperture. Experimental results under both simulations and darkroom measurements verify that the proposed method performs well on ISAR images of conic precession target with sparse aperture.

Sparse aperture inverse synthesis aperture radar (SA-ISAR) imaging is generally solved by compressed sensing (CS) methods or sparse signal recovery (SSR). Many SSR methods focus on the sparsity of radar images only, which achieves unsatisfactory results on structural data. In addition, most of the traditional CS algorithms suffer from a heavy computational burden. In this article, a new deep unfolding network called pattern-coupled sparse Bayesian learning (PCSBL)-generalized approximate message passing (GAMP)-Net is proposed. The proposed network structure can learn the model of block-sparse information from data to reconstruct images of better quality via fewer iteration steps. First, a complex-valued pattern-coupled hierarchical Gaussian prior model is established. Then, the GAMP algorithm is applied for computational Bayesian inference. Based on the previous PCSBL-GAMP framework, the iterative procedure is unrolled to be a deep network structure. A complex-valued backpropagation (BP) algorithm is derived for network training. Experiment results based on simulated and measured data validate the superiority of the proposed method over the traditional PCSBL-GAMP algorithm. Also, the proposed algorithm is ten times faster than the traditional PCSBL-GAMP algorithm.

This study proposes a model-driven deep network based on the linear alternating direction method of multipliers (L-ADMM), to solve the problem whereby the inverse synthetic aperture radar (ISAR) generates defocused images of targets exhibiting micro-motion with sparse aperture. The network unfolds the operation process of L-ADMM into a model-driven deep network, and automatically optimizes the parameters of the network through learning instead of manually adjusting the parameters, which can better obtain images. Analyses of data acquired through simulations and experimental measurements were used to compare the results of imaging obtained by L-ADMM-net with those of the range Doppler (R-D) algorithm, chirplet algorithm, and L-ADMM. The entropy of images obtained by L-ADMM-net was the lowest, and their image contrast and resolution were the highest. Moreover, L-ADMM-net can generate high-resolution images of targets exhibiting micro-motion with sparse aperture at a low signal-to-noise ratio (SNR), which verifies its robustness. It can also automatically update and adjust parameters more stably than L-ADMM. The proposed method significantly improves the resolution, robustness, and stability of images of targets exhibiting micro-motion in different situations compared with traditional methods, and can provide technical support for target recognition in the future.

Owing to the large degrees of freedom and reduced mutual coupling by producing difference coarrays, nonuniform linear arrays have aroused great interest in direction of arrival (DOA) estimation. Previous works have presented some new sparse arrays, such as the thinned coprime array. In this paper, we propose a generalized thinned coprime array by introducing the flexible inter-element spacings, where the conventional one can be seen as a special case. We derive closedform expression for the range of consecutive lags, written as the functions of the antenna numbers and inter-element spacings. We show that, after optimization, the proposed array can achieve more consecutive lags than the other coprime arrays. In particular, the optimized results also provide the minimum number of antenna pairs with small separation. Simulation results demonstrate the superiority of the proposed GTCA using the subspace-based method.

Diversity smoothing has been widely developed for angle estimation with bistatic multiple input multiple output (MIMO) radar in the presence of coherent targets, the parameter identifiability of which is an important issue. In this paper, we are devoted to establishing more accurate conditions by studying the positive definiteness of smoothed target covariance matrix. The antenna numbers of transmit and receive arrays are derived as functions of the target number and target structure. We show that the new results improve upon previous ones and recover them in special cases. Simulation results are presented that corroborate our theoretical findings.

Inverse synthetic aperture radar (ISAR) imaging for the target with micro-motion parts is influenced by the micro-Doppler (m-D) effects. In this case, the radar echo is generally decomposed into the components from the main body and micro-motion parts of target, respectively, to remove the m-D effects and derive a focused ISAR image of the main body. For the sparse aperture data, however, the radar echo is intentionally or occasionally under-sampled, which defocuses the ISAR image by introducing considerable interference, and deteriorates the performance of signal decomposition for the removal of m-D effects. To address this issue, this paper proposes a novel m-D effects removed sparse aperture ISAR (SA-ISAR) imaging algorithm. Note that during a short interval of ISAR imaging, the range profiles of the main body of target from different pulses are similar, resulting in a low-rank matrix of range profile sequence of main body. For the range profiles of the micro-motion parts, they either spread in different range cells or glint in a single range cell, which results in a sparse matrix of range profile sequence. From this perspective, the low-rank and sparse properties are utilized to decompose the range profiles of the main body and micro-motion parts, respectively. Moreover, the sparsity of ISAR image is also utilized as a constraint to eliminate the interference caused by sparse aperture. Hence, SA-ISAR imaging with the removal of m-D effects is modeled as a triply constrained underdetermined optimization problem. The alternating direction method of multipliers (ADMM) and linearized ADMM (L-ADMM) are further utilized to solve the problem with high efficiency. Experimental results based on both simulated and measured data validate the effectiveness of the proposed algorithm.

Inverse synthetic aperture radar (ISAR) imaging for the sparse aperture data is affected by considerable artifacts, because under-sampling of data produces high-level grating and side lobes. Noting the ISAR image generally exhibits strong sparsity, it is often obtained by sparse signal recovery (SSR) in case of sparse aperture. The image obtained by SSR, however, is often dominated by strong isolated scatterers, resulting in difficulty to recognize the structure of target. This paper proposes a novel approach to enhance the ISAR image obtained from the sparse aperture data. Although the scatterers of target are isolated in the ISAR image, they should be associated with the neighborhood to reflect some intrinsic structural information of the target. A convolutional reweighted l1minimization model, therefore, is proposed to model the structural sparsity of ISAR image. Specifically, the ISAR image is reconstructed by solving a sequence of reweighted l1problems, where the weight of each pixel used for the next iteration is calculated from the convolution of its neighbor values in the current solution. The problem is solved by the alternating direction of multipliers (ADMM) and linearized approximation, respectively, to improve the computational efficiency. Experimental results based on both simulated and measured data validate that the proposed algorithm is effective to enhance the ISAR image, robust to noise, and more impressively, very efficient to implement.

The micro-Doppler (m-D) effect caused by micro-motion degrades the readability of the inverse synthetic aperture radar (ISAR) image. To achieve well-focused ISAR image of the target with the micro-motion part, this paper proposes a novel approach for the removal of m-D effect of ISAR image. Note that the range profiles of the rigid body are similar to each other, making the respective data matrix low-rank. Those of the micro-motion part, in contrary, generally fluctuate in different range cells, whose data matrix is sparse. Therefore, the removal of m-D effect can be naturally solved by the robust principal component analysis (RPCA)-a convenient convex program to decompose an auxiliary matrix into a low-rank matrix and a sparse one. In RPCA, the rank of a matrix is described by the nuclear norm, which is convex but leads to a suboptimal solution. To address it, we utilize a nonconvex surrogate, i.e., the summation of logistic function of the singular values of a matrix, to approximate the rank. Moreover, the range profiles of the rigid body are generally locally similar. To capture this geometric structured information, we further introduce a Laplacian regularization into the model. Then, the Laplacian regularized nonconvex low-rank (LRNL) model is solved efficiently by the linearized alternating direction method (ADM). Extensive experimental results based on both simulated and measured data demonstrate the effectiveness of the proposed approach on the removal of m-D effect of ISAR image.

Formal program specifications are essential for various software engineering tasks, such as program verification, program synthesis, code debugging and software testing. However, manually inferring formal program specifications is not only time-consuming but also error-prone. In addition, it requires substantial expertise. Natural language comments contain rich semantics about behaviors of code, making it feasible to infer program specifications from comments. Inspired by this, we develop a tool, named C2S, to automate the specification synthesis task by translating natural language comments into formal program specifications. Our approach firstly constructs alignments between natural language word and specification tokens from existing comments and their corresponding specifications. Then for a given method comment, our approach assembles tokens that are associated with words in the comment from the alignments into specifications guided by specification syntax and the context of the target method. Our tool successfully synthesizes 1,145 specifications for 511 methods of 64 classes in 5 different projects, substantially outperforming the state-of-the-art. The generated specifications are also used to improve a number of software engineering tasks like static taint analysis, which demonstrates the high quality of the specifications.

In the case of sparse aperture, the coherence between pulses of radar echo is destroyed, which challenges inverse synthetic aperture radar (ISAR) autofocusing and imaging. Mathematically, reconstructing the ISAR image from the sparse aperture radar echo is a linear underdetermined inverse problem, which, by nature, can be solved by the fast developed compressive sensing (CS) or sparse signal recovery theory. However, the CS-based sparse aperture ISAR imaging algorithms are generally computationally heavy, which becomes the bottleneck of preventing their applications to the real-time ISAR imaging system. In this article, we propose a novel and computationally efficient ISAR autofocusing and imaging algorithm for sparse aperture. We first consider a generalized CS model for ISAR imaging and autofocusing with sparse and entropy-minimization regularizations, and then utilize the alternating direction method of multipliers (ADMM) algorithm to optimize the model. To improve computational efficiency, the matrix inversion is translated to an elementwise division with the usage of a partial Fourier dictionary, and the 2-D ISAR image is updated as a whole instead of range cellwise. To achieve autofocusing for sparse aperture, the phase error is estimated by minimizing the entropy of the ISAR image reconstructed in each iterative loop. Experiments based on both simulated and measured data validate that the proposed algorithm can achieve well-focused ISAR images within a few seconds, which is ten times faster than the reported sparse aperture ISAR imaging algorithms.

For sparse aperture (SA) radar imaging, the phase errors are difficult to be estimated, which challenges the traditional autofocusing for inverse synthetic aperture radar (ISAR) imaging. A novel Bayesian ISAR autofocusing algorithm for SA is proposed. We unfold the sparse Laplace prior to two layers so that the full variational Bayesian inference can be derived. To further exploit the prior knowledge on the structure of radar images, dependencies among adjacent pixels are considered to design a structured sparse prior. In addition, the minimum entropy criterion is utilized to estimate the phase error during the reconstruction of the ISAR image to achieve ISAR autofocusing. The superiority of the proposed method against the traditional sparsity-driven method is validated by the experimental results based on both simulated and measured data.

Sparse aperture ISAR autofocusing and imaging is generally achieved by methods of compressive sensing (CS), or, sparse signal recovery, because non-uniform sampling of sparse aperture disables fast Fourier transform (FFT)-the core of traditional ISAR imaging algorithms. Note that the CS based ISAR autofocusing methods are often computationally heavy to execute, which limits their applications in real-time ISAR systems. The improvement of computational efficiency of sparse aperture ISAR autofocusing is either necessary or at least highly desirable to promote their practical usage. This paper proposes an efficient sparse aperture ISAR autofocusing algorithm. To eliminate the effect of sparse aperture, the ISAR image is reconstructed by sparse Bayesian learning (SBL), and the phase error is estimated by minimum entropy during the reconstruction of ISAR image. However, the computation of expectation in SBL involves a matrix inversion with an intolerable computational complexity of at least O(L3). Here, in the Bayesian inference of SBL, we transform the time-consuming matrix inversion into an element-wise matrix division by the alternating direction method of multipliers (ADMM). An auxiliary variable is introduced to divide the computation of posterior into three simpler subproblems, bringing computational efficiency improvement. Experimental results based on both simulated and measured data validate the effectiveness as well as high efficiency of the proposed algorithm. It is 20-30 times faster than the SBL based sparse aperture ISAR autofocusing approach.

Obtained by wide band radar system, high resolution range profile (HRRP) is the projection of scatterers of target to the radar line-of-sight (LOS). HRRP reconstruction is unavoidable for inverse synthetic aperture radar (ISAR) imaging, and of particular usage for target recognition, especially in cases that the ISAR image of target is not able to be achieved. For the high-speed moving target, however, its HRRP is stretched by the high order phase error. To obtain well-focused HRRP, the phase error induced by target velocity should be compensated, utilizing either measured or estimated target velocity. Noting in case of under-sampled data, the traditional velocity estimation and HRRP reconstruction algorithms become invalid, a novel HRRP reconstruction of high-speed target for under-sampled data is proposed. The Laplacian scale mixture (LSM) is used as the sparse prior of HRRP, and the variational Bayesian inference is utilized to derive its posterior, so as to reconstruct it with high resolution from the under-sampled data. Additionally, during the reconstruction of HRRP, the target velocity is estimated via joint constraint of entropy minimization and sparseness of HRRP to compensate the high order phase error brought by the target velocity to concentrate HRRP. Experimental results based on both simulated and measured data validate the effectiveness of the proposed Bayesian HRRP reconstruction algorithm.

For sparse aperture (SA) radar echoes, the coherence between the undersampled pulses is destroyed, which challenges the effectiveness of the traditional autofocusing and scaling in inverse synthetic aperture radar (ISAR) imaging. A novel Bayesian ISAR autofocusing and scaling algorithm for sparse aperture is proposed, which utilizes Laplacian scale mixture, as the sparse prior of ISAR image, and variational Bayesian inference based on the Laplacian approximation to derive its posterior. In addition, it learns the phase error, rotational velocity, and center of target from radar echo automatically during the reconstruction of ISAR image, so as to achieve ISAR autofocusing and scaling for SA. Because the parameters learning is not easy to converge with the undersampled data, a modified Newton method based on joint constraint of entropy and sparsity is proposed to guarantee fast convergence in a right direction. Experimental results based on both simulated and measured data validate the robustness of the proposed ISAR imaging algorithm against SA and noise.

The cardinalized probability hypothesis density (CPHD) filter was proposed as a practical approximation to the multi-target Bayes filter with tractable computational complexity. However, the CPHD filter has limitations in dealing with missed detections, extracting target state in its particle implementations, and in maintaining track continuity. In this paper, a new improved CPHD filter is proposed as a solution to address these limitations, with efficient track continuity and extraction. This filter inherits tractable computational complexity and addresses the drawbacks of the standard CPHD filter. The proposed filter is implemented using Gaussian mixtures, and simulation results demonstrate the effectiveness of the proposed filter compared to the conventional multi-taraet filter in challenging scenarios.

User-centered philosophy is the trend of modern website design and its construction nowadays. In this paper, the methods of Kansei Engineering are imported to help improving the interact efficiency of an online Business-to-Business commercial website. First we collect the statistics result of up to 80 samples of user investigation, then the way of semantic differential has been used to reduce data dimension. The obtained representatively numerical knowledge makes it more effective for designer to carry interaction design in the prophase process. The website is redesigned accordingly. In later period, the skill of Kansei Engineering is still used to evaluate the effects of former design, new design and two competitors' website design. Experimental results show that the new design surpasses than others. And our proposed method could get more objective and more accurately of interaction than traditional experience-based design.

With the development of the very large scale integration circuit (VLSI), multi-core processors have been developing fast and provide a novel approach to meet the requirements of modern complex computing. Real-time SAR imaging is always a hard problem to deal with since the huge amount of data and complex processing. Therefore, we design a schedulable and scalable multi-core parallel architecture based on FPGA and map the fundamental Chirp Scaling algorithm to the system. The design of the master control core makes the system highly extensible and the dedicated computing cores are designed to accelerate the process. In addition, the system can meet the real-time requirements, and it performs well in the resource utilization and the FPGA chip takes up a small space as well.

This paper proposes a novel bistatic inverse synthetic aperture radar (ISAR) imaging algorithm for the target with complex motion under low signal to noise ratio (SNR) condition. Note the bistatic ISAR system generally suffers from a lower SNR than the monostatic one because of its non-mirror reflection geometry. A de-noising method, therefore, is proposed to improve SNR of range profiles, which accumulates the aligned range profiles non-coherently to obtain a window for noise suppression. Additionally, since the complex motion of target induces nonstationary Doppler, which is destructive to ISAR imaging, an optimal coherent processing interval (CPI) selection algorithm is further proposed to find out the interval where the Doppler is relatively stationary, so as to produce well-focused ISAR images. It utilizes the reassigned time-frequency (TF) method to obtain the high resolution instantaneous Doppler spectrum, and the minimum entropy criterion to select the optimal CPI, respectively. Note the selected CPI often contains too limited pulses to produce ISAR images with high resolution. A sparse aperture ISAR imaging method within the Bayesian framework is further proposed, which introduces the Laplacian scale mixture (LSM) model as the sparse prior, so as to reconstruct well-focused ISAR images with high resolution and low side lobes from the limited data. Compared with the traditional sparse Bayesian learning method, the proposed LSM based ISAR imaging performs superiorly on resolution improvement and noise reduction. Experimental results based on both simulated and measured data validate the effectiveness of the proposed algorithms.

In this paper, we have proposed a phase-derived-range (PDR) method for direct sampling of linear-frequency-modulated radar signals. This method is capable of measuring multiple scatterers' ranges of a target and yields root-mean-squared range error values at subwavelength level; therefore, it has great potential to measure micromotions of moving targets and is of significant importance for target recognition. The main challenge that we have solved is extracting ambiguous Doppler phases from high-resolution range profiles generated through a match filter. For guiding the implementation in radar systems, restraint conditions of radar parameters have been systematically justified based on the principle of ambiguity resolution. We have provided a systematic algorithm flowchart for PDR, especially for wideband direct sampling radars. Both simulated and experimental results under different radar parameter settings are presented and validate the performance.

Coherent compensation is a significant problem to multiband radar signal fusion. At present, the dominant root-MUSIC and linear least squares(LLS) based ways use band extrapolation (BWE) to generate the whole band signals, and then apply optimization to get the incoherent phases(ICPs). The precision of these methods are limited by the BWE error and the non-corresponding poles. Furthermore, the high dimensional optimization imposes a heavy burden on the computer. In order to improve the performance of the coherent compensation, the incoherent factors between sub-band signals are analyzed first, then in order to correct the non-corresponding poles, a poles correction process which makes use of minimum entropy principle is presented. Based on these, this paper provides an efficient way to estimate the ICPs of sub-band radar signals. Applications to simulated data verify that the proposed method can get better ICPs estimation results.

Abstract In the heterogeneous and unreliable channel environment of cognitive radio ad hoc networks (CRAHNs), a multipath route with channel assigned is preferable in both throughput and reliability. The cross‐layer multipath routing and channel assignment in CRAHNs is becoming a challenging issue. In this paper, this problem is characterized, formulated, and shown to be in the form of mixed integer programming. For this Non‐deterministic Polynomial‐time (NP)‐hard problem, the deficiency of the widely used linearization and sequential fixing algorithm is first analyzed. The main contribution of this paper is the development of a new backtracking algorithm with feasibility checking to search optimal solutions and a heuristic algorithm with high feasible solution‐obtained probability (HHFOP) for distributed application in CRAHNs. Through feasibility checking and solution bounds validating, backtracking algorithm with feasibility checking cuts off unnecessary searching space in early stage without loss of optimal solutions, making it much more efficient than brute searching. For practical application in CRAHNs with polynomial complexity, HHFOP first computes the maximal‐supported throughput through link‐channel assignment and link‐capacity coordination for each candidate path. Then the paths are combined, and the route throughput is optimized. Extensive simulation results demonstrate that HHFOP can achieve a high feasible solution‐obtained probability with little throughput degradation compared with linearization and sequential fixing algorithm, indicating its practicability for distributed applications in CRAHNs. Copyright © 2013 John Wiley & Sons, Ltd.

In this letter, a novel method for inverse synthetic aperture radar (ISAR) imaging under a low signal-to-noise ratio (SNR) condition is presented. The method is a preprocess of the range profiles before motion compensation and is based on the pseudomatched filter, whose impulse response is obtained by conjugating and reversing the average of the coarsely aligned range profile envelopes. With the utilization of the presented method, the SNR of the target range profiles is improved, the conventional ISAR motion compensation methods perform much better, and the ISAR image result is much better focused under a low SNR condition. Experimental results based on both the simulated and real data of an aircraft validate the performance of the presented method.

A novel method of precise radial range measurement based on wideband radar phase is presented in this paper for micromotion characteristic acquisition. The advantage of this method is the high precision with root-mean-square error values at subwavelength levels, while its difficulties are phase extraction from wideband radar echoes and resolution of ambiguous phase. After the analysis of the principle of radial range measurement based on radar phase, the method of extracting the Doppler phase from wideband radar echoes is proposed, following a comprehensive technical diagram for micromotion characteristic synthesis based on the wideband radar phase. Some restraint conditions for the resolution of the ambiguous phase are analyzed systematically according to the different characteristics of micromotion and radar parameters. The method provides a more precise tool to acquire the micromotion characteristic than the traditional Doppler frequency methods, and the experiments have shown the performances of wideband radar phase extraction and its ambiguity resolution.

Fast rotating targets such as gimbaled antennas or propeller blades may cause migration through resolution cells (MTRC) of the high-resolution range profile during the imaging time, which makes the inverse synthetic aperture radar image smeared. To solve this problem, a novel imaging method for fast rotating targets is proposed in this paper. The method is based on the segmental pseudo Keystone transform, which is designed to realize MTRC correction. The fast realization is achieved by employing the discrete match Fourier transform, which makes the algorithm feasible and simple. Experiments with simulated radar data demonstrate the performance of the proposed method.

In this article we propose techniques that enable efficient exploration of the 3D design space, where each logical block can span more than one silicon layer. Fine-grain 3D integration provides reduced intrablock wire delay as well as improved power consumption. However, the corresponding power and performance advantage is usually underutilized, since various implementations of multilayer blocks require novel physical design and microarchitecture infrastructure to explore 3D microarchitecture design space. We develop a cubic packing engine which can simultaneously optimize physical and architectural design for efficient vertical integration. This technique selects the individual unit designs from a set of single-layer or multilayer implementations to get the best microarchitectural design in terms of performance, temperature, or both. Our experimental results using a design driver of a high-performance superscalar processor show a 36% performance improvement over traditional 2D for 2--4 layers and 14% over 3D with single-layer unit implementations. Since thermal characteristics of 3D integrated circuits are among the main challenges, thermal-aware floorplanning and thermal via insertion techniques are employed to keep the peak temperatures below threshold.

Most previous 3D IC research focused on "stacking" traditional 2D silicon layers, so the interconnect reduction is limited to interblock delays. In this paper, we propose techniques that enable efficient exploration of the 3D design space where each logical block can span more than one silicon layers. Although further power and performance improvement is achievable through fine grain 3D integration, the necessary modeling and tool infrastructure has been mostly missing. We develop a cube packing engine which can simultaneously optimize physical and architectural design for effective utilization of 3D in terms of performance, area and temperature. Our experimental results using a design driver show 36% performance improvement (in BIPS) over 2D and 14% over 3D with single layer blocks. Additionally multi-layer blocks can provide up to 30% reduction in power dissipation compared to the single-layer alternatives. Peak temperature of the design is kept within limits as a result of thermal-aware floorplanning and thermal via insertion techniques.

In radar target imaging, motion induces a range-Doppler coupling effect, which results in distortion in synthesizing a range profile for a moving target. To eliminate or suppress the distortion of the synthetic range profile, conventional technology such as motion compensation requires velocity estimation. Unfortunately, for a high-speed moving target, it is difficult to achieve real-time accurate estimation of the velocity of the target. Based on phase cancellation, a new technology is proposed to achieve a motion target range profile via stepped-frequency waveform. The new technology does not need the estimation of the velocity. Hence, its computational cost can be minimized. It is also easier to use. The simulation result confirms the effect of the new technology

Energy dissipation from the issue queue and register file constitutes a large portion of the overall energy budget of an aggressive dynamically scheduled microprocessor. We propose techniques to save energy in these structures by reducing issue queue occupancy and by reducing unnecessary register file accesses that can result from speculative scheduling. Our results show a 44% reduction in issue queue occupancies and an 87% reduction in register file accesses for scheduling replays. Our data show that these savings can translate into a 52% saving in issue queue energy, a 13% savings in register file energy, and a 22% overall energy savings.

As future technologies push towards higher clock rates, traditional scheduling techniques that are based on wake-up and select from an instruction window fail to scale due to their circuit complexities. Speculative instruction schedulers can significantly reduce logic on the critical scheduling path, but can suffer from instruction misscheduling that can result in wasted issue opportunities.Misscheduled instructions can spawn other misscheduled instructions, only to be replayed over again and again until correctly scheduled. These "tornadoes" in the speculative scheduler are characterized by extremely low useful scheduling throughput and a high volume of wasted issue opportunities. The impact of tornadoes becomes even more severe when using Simultaneous Multithreading. Misschedulings from one thread can occupy a significant portion of the processor issue bandwidth, effectively starving other threads.In this paper, we propose Zephyr, an architecture that inhibits the formation of tornadoes. Zephyr makes use of existing load latency prediction techniques as well as coarse-grain FIFO queues to buffer instructions before entering scheduling queues. On average, we observe a 23% improvement in IPC performance, 60% reduction in hazards, 41% reduction in occupancy, and 48% reduction in the number of replays compared with a baseline scheduler.

In contemporary out-of-order superscalar design, high IPC is mainly achieved by exposing high instruction level parallelism (ILP). Scaling issue window size can certainly provide more ILP; however, future processor scaling demands threaten to limit the size of the issue window.In this study, we propose a dynamic instruction sorting mechanism that provides more ILP without increasing the size of the issue window. In our approach, early in the pipeline, we predict how long an instruction needs to wait before it can be issued, i.e. the waiting time for its operands to be produced. Using this knowledge, the instructions are placed into a sorting structure, which allows instructions with shorter waiting times enter the issue window ahead of those instructions with longer waiting times, preventing long-waiting instructions from clogging the issue queue.The accuracy in predicting instruction waiting times directly determines the effectiveness of our sorting mechanism. While most instructions have deterministic execution latencies, predicting load execution times is more difficult due to cache misses and in-flight loads. Loads are particularly challenging since their execution time can vary significantly. In this study, we examine techniques to predict load execution time accurately, based on data reference history.