Recent advances in real image super-resolution (Real-ISR) using diffusion models often inject low-resolution (LR) images via ControlNet, with training starting from the corresponding high-resolution (HR) images. This framework suffers from a training-inference mismatch: the noise scheduling scheme leaves residual signals even at the diffusion termination step, resulting in a non-zero signal-to-noise ratio (SNR). Consequently, the model learns to rely on residual high-frequency details from the HR input during training, details absent when inference begins from pure Gaussian noise, constraining texture fidelity. To bridge this gap, we propose a Dynamic Texture Noise Synthesizer (DTNS) that generates content-adaptive, spatially-varying texture noise from the LR image’s local gradient statistics, approximating missing high-frequency components. By embedding this synthesized noise into the initial pure noise, we provide an informed starting point for inference. A noise decoupling loss further steers synthesis away from irrelevant degradation patterns. Our method enhances perceptual quality and high-frequency reconstruction, outperforming existing techniques. This advancement is particularly beneficial for multimedia retrieval, where accurately enhancing low-quality query images can significantly improve retrieval precision and user experience.

Vision Foundation Models (VFMs) have demonstrated remarkable performance across various visual downstream tasks. To transfer their strong representation ability into lightweight CNNs, knowledge distillation serves as a promising approach. However, existing feature-based distillation methods typically perform alignment in spatial or channel dimensions, often neglecting the explicit semantic meaning of intermediate features. This semantic agnosticism leads to inefficient transfer, where students mimic the teacher’s activation patterns without grasping the underlying class-discriminative logic. In this paper, we propose Class-space Projection Distillation (CPD), a novel framework that enforces semantic alignment directly within the feature space. At the core of CPD is the Class-Projection Attention (CPA) module, which introduces a parallel dual-stream interaction mechanism: it complements standard spatial-texture retention with explicit class-semantic alignment. Specifically, CPA projects feature queries and keys into the global class space to generate Semantic Consistency Maps. Crucially, we design a versatile value projection strategy that leverages these maps for two distinct purposes: (1) Semantic-Guided Fusion, where values are projected to the feature dimension to refine student representations based on class consistency; and (2) Generative Class Activation Map (CAM) Supervision, where values are parallelly projected to the class dimension to directly generate intermediate CAMs, enforcing explicit pixel-level semantic constraints. Extensive experiments on CIFAR-100, Tiny-ImageNet, and ImageNet-1K demonstrate that CPD effectively bridges the architectural gap, consistently outperforming state-of-the-art methods.

The computational and memory demands of deep neural networks for vision tasks remain a critical barrier to their deployment on resource-constrained edge devices. Although knowledge distillation (KD) effectively transfers over-parameterized models’ knowledge into compact students, its efficacy diminishes substantially when a significant capacity gap exists between them. Current approaches often impose linear mapping constraints between output distributions, an assumption that becomes prohibitively restrictive under such capacity gaps. This paper proposes a fundamental relaxation of alignment requirements. Specifically, rather than enforcing strict parametric relationships, we experimentally validate that preserving monotonic rank correlation between teacher and student outputs suffices for effective knowledge transfer. To operationalize this insight, we introduceMonotonic Rank Knowledge Distillation, a novel framework that leverages differentiable approximations of Kendall’s rank correlation coefficient to measure and optimize rank-order consistency. Our methodology further decomposes rank correlation into inter-class and intra-class components, ensuring the student network retains both global discriminative patterns and fine-grained categorical distinctions inherent to the teacher’s outputs. Extensive experiments across CIFAR-100 and ImageNet-1K benchmarks validate the effectiveness of our approach, demonstrating consistent performance gains over state-of-the-art distillation methods. The proposed framework achieves superior generalization across diverse architectures, including CNN-based, MLP-based, and ViT-based, with particular efficacy in various compression scenarios.

Knowledge distillation has been proven to be an effective method for enhancing model performance, particularly in the domain of model compression. In this study, we propose a comprehensive approach that utilizes Brownian Distance Covariance (BDC) to measure the discrepancy between the logits produced by the teacher and student models. Unlike the conventional KL divergence used in traditional knowledge distillation, BDC captures not only linear relationships but also nonlinear dependencies, thereby overcoming the limitations of KL divergence and enabling the student model to learn more effectively from the teacher model. Additionally, our method aligns the discrepancies between the teacher and student models from both intra-class and inter-class perspectives. Extensive experiments demonstrate that our method achieves state-of-the-art (SOTA) performance across various network architectures and datasets. The code and resources related to this work are available at the following link: https://github.com/hengyin23654/BDCKD.

Post-training quantization (PTQ) has emerged as a promising approach for converting full-precision models into compact, low-precision models with minimal computational overhead, making them ideal for deployment in resource-constrained edge scenarios. While most existing PTQ techniques focus on minimizing the numerical discrepancy between model activations before and after quantization, such methods often overlook the inherent noise and distributional shifts caused by quantization, which can lead to severe performance degradation. To address this, we propose Distribution-Enhanced Reconstruction for PTQ (DEQuant), a novel approach that enhances the performance of quantized models by introducing a module that further enhances the alignment of activation pre- and post-quantization during model reconstruction. Extensive experiments demonstrate the effectiveness of DEQuant in several low-bit settings, achieving superior performance compared to existing methods. For instance, DEQuant achieves 14.18% accuracy on MobileNetV2 under the W2A2 configuration, representing a 5.72% improvement over the baseline QDrop and surpassing other baselines by 1–3%.

Deep neural networks for SAR image recognition often require compression for deployment on remote sensing platforms with limited computational and storage resources. Knowledge distillation (KD) is a key approach to improving the accuracy of lightweight networks. However, existing KD methods face challenges when applied to SAR images due to the small dataset size and the high noise in SAR images. To address this, this paper proposes a novel knowledge distillation method that relaxes the requirement for a strict linear relationship between the outputs of lightweight and large models, focusing instead on maintaining a Monotonic Relationship (MRKD). This reduces the difficulty of the KD task. Experiments on various SAR image classification and object detection datasets demonstrate that MRKD achieves state-of-the-art performance improvements for lightweight networks.

The Zero-Shot Sketch-Based Image Retrieval (ZS-SBIR) task seeks to match images with the same semantic essence as a hand-drawn sketch from a vast image repository. Given the stark contrast in information density between simple-line sketches and detailed images, this task encounters two formidable challenges: 1) Network layers focus differently on semantically relevant features across the two domains, and 2) The sparse information in sketches hampers the extraction of meaningful features. In response, we introduce the innovative Cross-Domain Feature Semantic Calibration (CD-FSC) model. This model begins by evaluating semantic correlations between domains and layers through attention map analyses in vision transformers to ensure precise semantic alignment. Subsequently, it harnesses category associations learned from the image domain to bolster semantic learning in the sketch domain. Our extensive comparative experiments across three prevalent ZS-SBIR datasets affirm that our model sets a new benchmark, outperforming current leading methods.

Recently, computation offloading methods have greatly improved the Quality of Experience (QoE) in Multi-access Edge Computing (MEC) by offloading tasks to the edge servers. Since well-coordinated actions of Terminal Devices (TDs) are critical to improving the performance of the entire individual system, many practical MEC-based applications, i.e., firefighting robots and unmanned aerial vehicles, require great teamwork among TDs. However, real-world scenarios are usually bound by resource conditions. For instance, network connectivity may weaken or experience interruptions during emergency situations. In cases where the communication medium is utilized by multiple TDs, achieving effective coordination poses a significant challenge. In this paper, we propose a computation offloading scheme based on Scheduled Multi-agent Deep Reinforcement Learning (SMDRL) to make the most efficient decision in a resource-constrained scenario. First, we design a virtual energy queue based on the MEC system and maximize the QoE (related to service delay and energy consumption) in a real-time manner. Subsequently, we propose a scheduled multi-agent deep reinforcement learning algorithm to support each TD in learning how to encode messages, select actions, and schedule itself based on the received messages. Furthermore, a TopK mechanism is introduced. This mechanism chooses the most crucial TDs to broadcast their messages, and then the computation offloading problem in a communication-constrained MEC environment can be solved in a low-communication manner. Also, we prove that even under limited communication conditions, our proposed methods can still lead to the close-to-optimal performance. The final performance analysis shows that the developed scheme has significant advantages over other representative schemes.

In recent years, deep convolutional neural networks (DCNNs) have been widely used in the task of ship target detection in synthetic aperture radar (SAR) imagery. However, the vast storage and computational cost of DCNN limits its application to spaceborne or airborne onboard devices with limited resources. In this paper, a set of lightweight detection networks for SAR ship target detection are proposed. To obtain these lightweight networks, this paper designs a network structure optimization algorithm based on the multi-objective firefly algorithm (termed NOFA). In our design, the NOFA algorithm encodes the filters of a well-performing ship target detection network into a list of probabilities, which will determine whether the lightweight network will inherit the corresponding filter structure and parameters. After that, the multi-objective firefly optimization algorithm (MFA) continuously optimizes the probability list and finally outputs a set of lightweight network encodings that can meet the different needs of the trade-off between detection network precision and size. Finally, the network pruning technology transforms the encoding that meets the task requirements into a lightweight ship target detection network. The experiments on SSDD and SDCD datasets prove that the method proposed in this paper can provide more flexible and lighter detection networks than traditional detection networks.

The huge computing and storage requirements of deep con-volutional neural networks (DCNNs) limit their application on edge computing devices. In this article, we propose an attention mechanism based on the feature map quality evaluation algorithm (IQE). The knowledge distillation method based on the IQE attention mechanism uses the IQE method to identify important knowledge in the pre-trained SAR target recognition deep neural network. Then in the process of knowledge distillation, the lightweight network is forced to focus on the learning of important knowledge. Through this mechanism, the method proposed in this paper can efficiently transfer the knowledge of the pre-trained SAR target recognition network to the lightweight network, which makes it is possible to deploy the SAR target recognition algorithm on the edge computing platform. Comparison experiments with several commonly used knowledge distillation methods have proved the effectiveness of our proposed method. In addition, we also verified the performance of the lightweight network obtained by our method on the edge platform based on the K210 processor.

With the maturity of synthetic aperture radar (SAR) technology, the problem of imbalanced data has gradually emerged. This problem makes it difficult for the automatic target recognition (ATR) model to properly learn the classification boundaries of majority and minority category target samples. In this article, we propose an ATR model with new architecture, called the cost-sensitive awareness-based automatic target recognition (CA-ATR) model, which provides an effective way of solving the problem of imbalanced data. Aimed at the two issues caused by imbalanced data on ATR models, the proposed method solves the problems from both the data and algorithm levels. At the data level, CA-ATR avoids adverse correlations among the target samples through different oversampling methods. By making the ATR model cost-sensitive, the proposed method also avoids the empirical risk preference of the ATR model for majority category target samples at the algorithm-level. At the same time, CA-ATR can autonomously learn different cost-sensitive awareness from different imbalanced data sets. The awareness enables the ATR model to more accurately learn the classification boundaries between target samples that belong in different categories. Several experimental results show the superiority of the proposed approach based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) data set. Compared with other imbalanced learning methods, the proposed method is able to solve different types of imbalanced data problems.

Since there are differences in the natural frequency of various synthetic aperture radar (SAR) target samples in reality, the problem of imbalanced data on the automatic target recognition (ATR) model has gradually appeared in recent years. The problem makes the classification boundary learned by the ATR model often fuzzy or even wrong. In this article, an SAR target sample generation method was proposed, called demand-driven generative adversarial nets (DDGANs), which provided an effective way to implement imbalanced data learning. When the imbalanced data exacerbated the deterioration of the minority category target samples distribution, the proposed method generated samples to alleviate this negative impact. The proposed method innovatively used two convolutional neural networks to form the discriminator of DDGAN. Among them, a convolutional neural network was used to determine whether the generated sample is real or fake. Moreover, another convolutional neural network can simultaneously dig out the generation demands of different categories of target samples when recognizing the generated samples. The generation demands enabled DDGAN to allocate different generation capabilities to different target samples on demand, thereby alleviating the negative impact of data imbalance. At the same time, DDGAN can autonomously learn the generation demands from imbalanced training sets. Several experimental results based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset showed the advantages of DDGAN. Compared with existing imbalanced learning algorithms, the proposed method had obvious superiority in recognition performance and data generation efficiency.

The rapid development of generative adversarial nets (GANs) has led to an increasing number of applications for the synthetic aperture radar (SAR) automatic target recognition (A-TR) with a small sample set in the past few years. However, the generated samples by the GAN s sometimes even lead to a decrease in the performance of the ATR model. In this paper, we propose a filtering approach to address this harm of generated samples. The proposed filtering approach is based on a stable generation model. The stable generation model can continuously and stably generate different batches of target samples. Then, multiple SVMs trained by different SAR target sample sets provide pseudo-labels to the other SVMs to improve the accuracy of the filtering results. Therefore, the proposed approach improves the recognition ability of the A-TR model dynamically while continuously filtering generated target samples. Several experimental results show the superiority of the proposed filtering approach based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) data set. When the number of training samples is 14.5% of the original training set, the recognition rate of the ATR model still reaches 91.27% with the help of the proposed approach.

Deep convolutional neural networks (DCNNs) have been widely used for SAR image target recognition. However, the huge demands of DCNNs for computing, storage, and energy resources limit their use on edge computing devices. In this article, we propose a method based on image quality evaluation (IQE) criterion to prune deep neural networks. We use IQE criterion to identify unimportant filters, and then remove them, to obtain a lightweight network while maintaining the performance of the neural network as much as possible. Besides, we verified the effectiveness of our method on the MSTAR dataset with cheap edge computing devices.