Jun Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jun Yang has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Congyuan Xu
College of Artificial Intelligence, Jiaxing University, Jiaxing 314001, China
3 shared publications - Donghui Zhang
School of Control Engineering, Northeastern University, Shenyang 110004, China
2 shared publications - Le Yang
School of Intelligent Sensing and Optoelectronic Engineering, Northeastern University, Qinhuangdao 066004, China
2 shared publications Mingfeng Jiang
2 shared publicationsZihao Liu
College of Information Science and Engineering, Jiaxing University, Jiaxing, China
2 shared publicationsCongyuan Xu
1 shared publication
Latest Publications
- Open Access
- 1 April 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (7) , 1175
Frequency-Domain Feature Learning Network for Joint Image Demosaicing and Denoising
The methods employed for image demosaicing and denoising play a pivotal role in image acquisition and restoration, and have been extensively studied over the past few decades. Traditionally, these tasks are performed sequentially, with demosaicing followed by denoising, or vice versa, treating each process independently. While this approach can enhance image quality, it often leads to issues such as color inaccuracies and information loss, as the outcome of the first task influences the second. Consequently, the integration of joint demosaicing and denoising (JDD) has become a focal point in recent research. Deep convolutional neural networks have shown promising results in addressing JDD challenges. This study introduces an end-to-end network, termed the Frequency-domain Features learning Network (FFNet), designed to tackle the JDD problem. Unlike conventional methods that focus on spatial domain features, FFNet utilizes frequency-domain (FD) characteristics to capture both global and local image details. Based on the vision Transformer architecture, FFNet consists of two key components: a global Fourier block (GFB), which uses global attention to determine the weights of FD parameters, and an MLP-based local Fourier block (LFB), which improves local feature extraction. These blocks are integrated with a channel attention mechanism to form the frequency-domain attention block (FAB), the core element of FFNet. Extensive experimental results on benchmark datasets demonstrate that FFNet achieves superior performance in terms of both quantitative metrics (PSNR/SSIM) and visual quality compared to existing state-of-the-art JDD methods. Furthermore, we provide a comprehensive analysis of its computational efficiency, including parameter count, FLOPs, and inference time, showing a competitive trade-off between performance and complexity.
- 1 April 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (7) , 1175
Open AccessFrequency-Domain Feature Learning Network for Joint Image Demosaicing and Denoising
The methods employed for image demosaicing and denoising play a pivotal role in image acquisition and restoration, and have been extensively studied over the past few decades. Traditionally, these tasks are performed sequentially, with demosaicing followed by denoising, or vice versa, treating each process independently. While this approach can enhance image quality, it often leads to issues such as color inaccuracies and information loss, as the outcome of the first task influences the second. Consequently, the integration of joint demosaicing and denoising (JDD) has become a focal point in recent research. Deep convolutional neural networks have shown promising results in addressing JDD challenges. This study introduces an end-to-end network, termed the Frequency-domain Features learning Network (FFNet), designed to tackle the JDD problem. Unlike conventional methods that focus on spatial domain features, FFNet utilizes frequency-domain (FD) characteristics to capture both global and local image details. Based on the vision Transformer architecture, FFNet consists of two key components: a global Fourier block (GFB), which uses global attention to determine the weights of FD parameters, and an MLP-based local Fourier block (LFB), which improves local feature extraction. These blocks are integrated with a channel attention mechanism to form the frequency-domain attention block (FAB), the core element of FFNet. Extensive experimental results on benchmark datasets demonstrate that FFNet achieves superior performance in terms of both quantitative metrics (PSNR/SSIM) and visual quality compared to existing state-of-the-art JDD methods. Furthermore, we provide a comprehensive analysis of its computational efficiency, including parameter count, FLOPs, and inference time, showing a competitive trade-off between performance and complexity.
- Open Access
- 1 April 2026
- Published by MDPI AG in Electronics
- Vol. 15 (7) , 1462
Image Colorization with Residual Attention U-Net
Image colorization aims to add plausible colors to grayscale images. However, existing methods often suffer from detail loss, dull colors, and unrealistic results. To address these issues, we propose a novel image colorization method based on a residual attention U-Net. First, a shallow feature extraction module with a fusion attention mechanism is designed to capture shallow features. Second, a residual attention U-Net is constructed by integrating a residual attention module into an improved U-Net architecture. Finally, we fuse the extracted shallow features with the shallow attention features within the residual attention U-Net to enhance detail preservation and improve colorization quality. Experimental results on the summer2winter dataset show that our method improves the average PSNR by 1.32 dB and SSIM by 0.0139, while reducing LPIPS by 0.01. Furthermore, our method achieves the best average PSNR and LPIPS on the NCData and COCO-Stuff datasets. Visual results demonstrate that our approach preserves fine details, produces more vibrant colors, and achieves a higher degree of realism and naturalness.
- 1 April 2026
- Published by MDPI AG in Electronics
- Vol. 15 (7) , 1462
Open AccessImage Colorization with Residual Attention U-Net
Image colorization aims to add plausible colors to grayscale images. However, existing methods often suffer from detail loss, dull colors, and unrealistic results. To address these issues, we propose a novel image colorization method based on a residual attention U-Net. First, a shallow feature extraction module with a fusion attention mechanism is designed to capture shallow features. Second, a residual attention U-Net is constructed by integrating a residual attention module into an improved U-Net architecture. Finally, we fuse the extracted shallow features with the shallow attention features within the residual attention U-Net to enhance detail preservation and improve colorization quality. Experimental results on the summer2winter dataset show that our method improves the average PSNR by 1.32 dB and SSIM by 0.0139, while reducing LPIPS by 0.01. Furthermore, our method achieves the best average PSNR and LPIPS on the NCData and COCO-Stuff datasets. Visual results demonstrate that our approach preserves fine details, produces more vibrant colors, and achieves a higher degree of realism and naturalness.
- Open Access
DELP-Net: A Differentiable Entropy Layer Pyramid Network for End-to-End Low-Rate DoS Detection
Low-rate Denial-of-Service (LDoS) attacks exploit periodic traffic pulses to trigger congestion while maintaining a low average rate, making them highly stealthy and difficult to distinguish from legitimate bursty traffic using threshold-based or simple statistical detectors. To address this challenge, this paper proposes DELP-Net, an end-to-end Differentiable Entropy Layer Pyramid Network for window-level online LDoS detection directly from raw traffic. DELP-Net combines a multi-scale one-dimensional convolutional pyramid with a differentiable Rényi-entropy-driven attention mechanism to capture distributional regularity and weak repetitive patterns characteristic of LDoS traffic. In addition, an entropy-conditioned temporal convolutional network is employed to model cross-window periodic dependencies in a lightweight manner, together with an entropy-regularized hybrid loss to enhance robustness under complex background traffic. Experiments on the low-rate DoS dataset show that DELP-Net achieves an average F1 score of 0.9877 across six LDoS attack types, with a detection rate of 98.69% and a false-positive rate of 1.15%, demonstrating its effectiveness and suitability for practical online intrusion detection deployments.
Open AccessDELP-Net: A Differentiable Entropy Layer Pyramid Network for End-to-End Low-Rate DoS Detection
Low-rate Denial-of-Service (LDoS) attacks exploit periodic traffic pulses to trigger congestion while maintaining a low average rate, making them highly stealthy and difficult to distinguish from legitimate bursty traffic using threshold-based or simple statistical detectors. To address this challenge, this paper proposes DELP-Net, an end-to-end Differentiable Entropy Layer Pyramid Network for window-level online LDoS detection directly from raw traffic. DELP-Net combines a multi-scale one-dimensional convolutional pyramid with a differentiable Rényi-entropy-driven attention mechanism to capture distributional regularity and weak repetitive patterns characteristic of LDoS traffic. In addition, an entropy-conditioned temporal convolutional network is employed to model cross-window periodic dependencies in a lightweight manner, together with an entropy-regularized hybrid loss to enhance robustness under complex background traffic. Experiments on the low-rate DoS dataset show that DELP-Net achieves an average F1 score of 0.9877 across six LDoS attack types, with a detection rate of 98.69% and a false-positive rate of 1.15%, demonstrating its effectiveness and suitability for practical online intrusion detection deployments.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 33, 2390-2394
TFS-Net: Time–Frequency Signal Modeling for Few-Shot Low-Rate DoS Detection
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 33, 2390-2394
TFS-Net: Time–Frequency Signal Modeling for Few-Shot Low-Rate DoS Detection
- 2 March 2025
- Published by SAGE Publications in The European Journal on Artificial Intelligence
- Vol. 38 (2) , 238-257
MSNet: Food Image Classification Model Based on Lightweight Convolutional Neural Network
With the advancement of technology, the demand for healthy eating has increased, making food classification a research hotspot. Existing deep learning-based food image classification models demonstrate high accuracy but require substantial computational resources, limiting their use on resource-constrained devices. In this study, a lightweight convolutional neural network model named MSNet for food classification is proposed. MSNet mainly consists of M Blocks and S Blocks. The M Block uses improved depthwise convolution to reduce the computational cost of conventional convolutions, and the S Block uses channel shuffle techniques to enhance feature information flow between channels without increasing additional computation, effectively capturing relationships between different channel features. Experimental results on three benchmark datasets (ETHZ Food-101, Vireo Food-172, and ISIA Food-500) show that MSNet achieves top-1 accuracies of 86.24%, 87.98%, and 65.70%, with model sizes of 13.8 MB, 15.9 MB, and 25.4 MB, respectively, outperforming mainstream models in terms of computational efficiency. Further quantization produces two MSNet-Lite variants with competitive model size, while maintaining high accuracy and significantly improving inference speed. Additionally, visualization analysis indicates that MSNet effectively extracts essential features of food images, offering good interpretability and generalization across datasets of varying complexity. The proposed MSNet model provides a feasible solution for practical deployment in food classification tasks on mobile and embedded devices.
- 2 March 2025
- Published by SAGE Publications in The European Journal on Artificial Intelligence
- Vol. 38 (2) , 238-257
MSNet: Food Image Classification Model Based on Lightweight Convolutional Neural Network
With the advancement of technology, the demand for healthy eating has increased, making food classification a research hotspot. Existing deep learning-based food image classification models demonstrate high accuracy but require substantial computational resources, limiting their use on resource-constrained devices. In this study, a lightweight convolutional neural network model named MSNet for food classification is proposed. MSNet mainly consists of M Blocks and S Blocks. The M Block uses improved depthwise convolution to reduce the computational cost of conventional convolutions, and the S Block uses channel shuffle techniques to enhance feature information flow between channels without increasing additional computation, effectively capturing relationships between different channel features. Experimental results on three benchmark datasets (ETHZ Food-101, Vireo Food-172, and ISIA Food-500) show that MSNet achieves top-1 accuracies of 86.24%, 87.98%, and 65.70%, with model sizes of 13.8 MB, 15.9 MB, and 25.4 MB, respectively, outperforming mainstream models in terms of computational efficiency. Further quantization produces two MSNet-Lite variants with competitive model size, while maintaining high accuracy and significantly improving inference speed. Additionally, visualization analysis indicates that MSNet effectively extracts essential features of food images, offering good interpretability and generalization across datasets of varying complexity. The proposed MSNet model provides a feasible solution for practical deployment in food classification tasks on mobile and embedded devices.
- Open Access
- 1 January 2023
- Published by National Library of Serbia in Computer Science and Information Systems
- Vol. 20 (2) , 813-829
Inverse halftoning based on sparse representation with boosted dictionary
Under the imbalanced dataset, the performance of the base-classifier, the computing method of weight of base-classifier and the selection method of the base-classifier have a great impact on the performance of the ensemble classifier. In order to solve above problem to improve the generalization performance of ensemble classifier, a selective ensemble learning algorithm based on under-sampling for imbalanced dataset is proposed. First, the proposed algorithm calculates the number K of under-sampling samples according to the relationship between class sample density. Then, we use the improved K-means clustering algorithm to under-sample the majority class samples and obtain K cluster centers. Then, all cluster centers (or the sample of the nearest cluster center) are regarded as new majority samples to construct a new balanced training subset combine with the minority class?s samples. Repeat those processes to generate multiple training subsets and get multiple base-classifiers. However, with the increasing of iterations, the number of base-classifiers increase, and the similarity among the base-classifiers will also increase. Therefore, it is necessary to select some base-classifier with good classification performance and large difference for ensemble. In the stage of selecting base-classifiers, according to the difference and performance of base-classifiers, we use the idea of maximum correlation and minimum redundancy to select base-classifiers. In the ensemble stage, G-mean or F-mean is selected to evaluate the classification performance of base-classifier for imbalanced dataset. That is to say, it is selected to compute the weight of each base-classifier. And then the weighted voting method is used for ensemble. Finally, the simulation results on the artificial dataset, UCI dataset and KDDCUP dataset show that the algorithm has good generalization performance on imbalanced dataset, especially on the dataset with high imbalance degree.
- 1 January 2023
- Published by National Library of Serbia in Computer Science and Information Systems
- Vol. 20 (2) , 813-829
Open AccessInverse halftoning based on sparse representation with boosted dictionary
Under the imbalanced dataset, the performance of the base-classifier, the computing method of weight of base-classifier and the selection method of the base-classifier have a great impact on the performance of the ensemble classifier. In order to solve above problem to improve the generalization performance of ensemble classifier, a selective ensemble learning algorithm based on under-sampling for imbalanced dataset is proposed. First, the proposed algorithm calculates the number K of under-sampling samples according to the relationship between class sample density. Then, we use the improved K-means clustering algorithm to under-sample the majority class samples and obtain K cluster centers. Then, all cluster centers (or the sample of the nearest cluster center) are regarded as new majority samples to construct a new balanced training subset combine with the minority class?s samples. Repeat those processes to generate multiple training subsets and get multiple base-classifiers. However, with the increasing of iterations, the number of base-classifiers increase, and the similarity among the base-classifiers will also increase. Therefore, it is necessary to select some base-classifier with good classification performance and large difference for ensemble. In the stage of selecting base-classifiers, according to the difference and performance of base-classifiers, we use the idea of maximum correlation and minimum redundancy to select base-classifiers. In the ensemble stage, G-mean or F-mean is selected to evaluate the classification performance of base-classifier for imbalanced dataset. That is to say, it is selected to compute the weight of each base-classifier. And then the weighted voting method is used for ensemble. Finally, the simulation results on the artificial dataset, UCI dataset and KDDCUP dataset show that the algorithm has good generalization performance on imbalanced dataset, especially on the dataset with high imbalance degree.