检索结果-内蒙古大学图书馆

Contour-texture preservation transformer for face super-resolution

NEUROCOMPUTING 2025年 626卷

作者： Wu, Ziyi Zhang, Yanduo Lu, Tao Zhao, Kanghui Wang, Jiaming Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan Peoples R China

In the field of computer vision, face super-resolution (FSR) technology is an important tool for enhancing the performance of basic tasks such as face recognition and video surveillance. However, when faced with complex face images, existing FSR methods often rely on the Transformer model to improve image quality through its powerful global modeling capabilities. Yet, they tend to be slightly insufficient in local feature extraction due to a lack of adequate local detail capture capabilities. To alleviate these problems, we propose a novel contour- texture preservation Transformer (CTP) method for FSR. This method consists of two key components: the multi-scale attention enhancement block (MSAEB), which captures and fuses image features of different scales to improve the detail level of feature representation, and provides high-quality input for the contour-texture Transformer enhancement block (CTTEB). Additionally, CTTEB integrates convolution operations to enhance local feature extraction and improve feature expression. The feed-forward network (FFN) we introduced ensures the full fusion of global and local information. By combining MSAEB and CTTEB into a residual progressive attention group (RPAG), the network gradually extracts and fuses multi-scale features, ultimately achieving dual preservation of contour structure and texture details. Experiments show that our method achieves the best results on LFW, FFHQ, CelebA, and Helen, with a 0.32 dB increase in PSNR, a 0.0126 increase in SSIM, and a 0.0056 increase in FSIM over the second-best model. Experiments on real-world datasets SCface and Chokepoint confirm that the CTP method excels in both FSR reconstruction and face recognition, verifying its effectiveness.

关键词： Face super-resolution Contour structure Texture details Face recognition

来源：评论

学校读者我要写书评

暂无评论

Action quality assessment via moment aware network

引用

EVOLVING SYSTEMS 2025年第2期16卷 1-10页

作者： Han, Jifeng Zhang, Yanduo Lu, Tao Wang, Jiaming Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan Peoples R China

Recent action quality assessment methods attempt to decompose the entire video into sub-actions of the same level, which still suffers from information redundancy and lacks certain interpretability. Through studying and analysing the evaluation rules of relevant sports events, we assume that certain critical moments in the actions will significantly impact the assessment of the entire action, if the weight of these critical moments in the evaluation process can be increased, then the results can be more accurate and more persuasive. To validate this idea, we propose a moment-aware approach for action quality assessment, a novel moment-aware module is used to explore the key moments in action globally and locally to obtain more reliable evaluation results. Extensive experiments have shown that our module captures key moments that are close to the focus of human referees, and our method has also achieved favorable results on multiple public AQA benchmarks, the Spearman's rank correlation reached 0.9360 on FineDiving dataset and 0.8797 on AQA-7 dataset.

关键词： Action Quality Assessment Fine-Grained Action Analyse Temporal Pattern of Sports

来源：评论

学校读者我要写书评

暂无评论

RO-BEV: Towards Robust BEV Feature Enhancement for Surround-View Instance Prediction

引用

ELECTRONICS LETTERS 2025年第1期61卷

作者： Yang, Zhi Lu, Tao Wang, Jiaming Lang, Xiujuan Fu, Shichang Han, Jifeng Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan Peoples R China

The most existing instance prediction methods often lack practical understanding and enhancement of bird's-eye view (BEV) spatial-temporal features, particularly distant features. To alleviate the above problems, we propose a robust BEV feature enhancement network (RO-BEV). The proposed RO-BEV includes two major components: the novel adaptive linearly increasing dividing (ALID) strategy and the hierarchical cross-scale fusion (HCSF) module. Specifically, the ALID strategy can mitigate instance loss caused by insufficient features by providing a linearly increasing 3D representation to build a robust BEV feature. Then, the proposed HCSF module enhances spatial-temporal BEV features in the latent space to predict vehicle instances by sampling from feature distributions at different scales and timestamps. Experimental results on the nuScenes dataset show that our RO-BEV outperforms existing state-of-the-art methods.

关键词： autonomous driving Bird's-eye view computer vision image processing instance prediction

来源：评论

学校读者我要写书评

暂无评论

WA-Net: Wavelet Integrated Attention Network for Silk and Bamboo character recognition

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2025年 140卷

作者： Li, Shengnan Zhou, Chi Wang, Kaili Wuhan Inst Technol Sch Comp Sci & Engn Hubei Key Lab Intelligent Robot Wuhan 430205 Peoples R China

Chu Bamboo and Silk ancient Chinese character (CBSC) was originated in the Chu state over 2000 years ago, representing an intermediate script between oracle bone script and seal script. Existing text images have degraded and suffered damage due to their ancient historical origins and insufficient preservation. Due to distinct structural and stroke texture characteristics, significant differences exist between CBSC and contemporary characters, posing challenges for intelligent recognition. Targeting these aforementioned characteristics, we propose a method called Wavelet Integrated Attention Network (WA-Net). This method integrates discrete wavelet transform and attention mechanisms to extract more discriminative features from severe noise interference and degraded text images. Additionally, a dataset named Chu Bamboo and Silk 730 (Chu730) for CBSC recognition has been created due to the lack of publicly available datasets. WA-Net introduces the discrete wavelet attention among layer (L-DWT) to broaden the feature learning space of convolutional neural networks into the wavelet domain, capturing latent information across various frequencies. Subsequently, a wavelet convolution (C-DWT) module is proposed to mitigate the partial information loss of conventional convolution operations. In the W-bneck module, the SE (Squeeze-and-Excitation) attention module and average pooling downsampling are introduced to enhance the extraction of valuable feature maps. Extensive experiments were conducted, including a baseline method that achieved top-1 recognition accuracy of 87.42%. The proposed method achieved an accuracy of 89.27%, and other top-n results also significantly surpassed the baseline accuracy. Other experiment results demonstrate the superiority of the proposed modules and theirvaluable applications in ancient text intelligent recognition and cultural heritage digital preservation. Furthermore, this approach holds significant promise in facilitating the study of other

关键词： Chu Bamboo and Silk ancient Chinese character Text recognition Discrete wavelet transform Wavelet attention

来源：评论

学校读者我要写书评

暂无评论

Prompt learning and multi-scale attention for infrared and visible image fusion

引用

INFRARED PHYSICS & TECHNOLOGY 2025年 145卷

作者： Li, Yanan Ji, Qingtao Jiao, Shaokang Key Lab Crop Prod & Smart Agr Yunnan Prov Kunming 650201 Yunnan Peoples R China Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan 450205 Hubei Peoples R China

Sugar beet is the world's second largest sugar crop. In modern sugar beet farms, high-quality images serve as the foundation and guarantee for crop yield prediction, weed identification, automated harvesting and other farm management tasks. Visible images are sensitive to changes in lighting conditions, while infrared images lack sufficient texture details. Neither modality can individually provide high-quality images. Fusing infrared and visible images can simply and effectively improve farm image quality. Since the original images in the fusion task usually contain noise, direct fusion of noisy images will affect the feature recognition ability of the network. Therefore, this paper proposes a fusion model based on prompt learning and multi-scale attention mechanism, termed Prompt Denoising Adaptive Fusion (PDAFusion). The model introduces a Prompt Denoising Block (PDB) in the encoder, which dynamically removes modality noise through learnable prompts during the feature encoding stage. Additionally, a Multi-Scale Adaptive Fusion Block (MSAFB) is designed, which utilizes attention mechanisms to compute the relative importance of features at different scales, adaptively allocating modality weights to enhance the quality of the fused image. The effectiveness of the proposed PDBFusion has been demonstrated on the Sugar Beets 2016 dataset, and the robustness of the model was validated on datasets such as RoadScene and TNO.

关键词： Prompt learning Multi-scale attention Multimodal Multispectral fusion Image denoising

来源：评论

学校读者我要写书评

暂无评论

Two-stage learning framework for single image deraining

引用

IET IMAGE PROCESSING 2023年第5期17卷 1449-1463页

作者： Jiang, Rui Li, Yaoshun Chen, Cheng Liu, Wei Wuhan Inst Technol Sch Comp Sci & Engn Hubei Key Lab Intelligent Robot Wuhan Peoples R China Wuhan Inst Technol Sch Comp Sci & Engn Hubei Key Lab Intelligent Robot Wuhan 430205 Peoples R China

Single image deraining methods have been extensively studied for its ability to remarkably improve the performance of computer vision tasks in rainy environments. However, most existing rain removal methods still have two major drawbacks which are hindering the technology development. First, the rain streaks are seriously coupled with the background information in a single rainy image, which leads to incorrect identification of rain streaks by many methods and further makes the loss of texture details in the rain removal results. Second, they spend excessive computational cost, which is not conducive to practical applications. To address these issues, a progressive separation network (PSN) is proposed by decomposing the rain removal task into two stages, the bilateral grid learning stage and the joint feature refinement stage, from a novel perspective. The bilateral grid learning stage is designed to expand the distance between the rain streaks and the background information while preserving the image edge details to guide the subsequent refinement. For the joint feature refinement stage, a dual-path interaction module is constructed to dynamically and gradually decouple the rain streak content and the intermediate features of the clear image details. In addition, an activation-free feature refinement block is designed to further improve the computational efficiency by removing or replacing the activation function without loss of accuracy. Extensive experiments on synthetic and real datasets show that PSN outperforms state-of-the-art rain removal methods in terms of quantitative accuracy and subjective visual quality. Furthermore, competitive results are derived by extending PSN to the defogging task.

关键词： bilateral grid joint feature refinement progressive separation

来源：评论

学校读者我要写书评

暂无评论

MLAD: Manifest and latent anomaly detection based on the integration of reconstruction and MLFP-KNN methods

引用

MEASUREMENT SCIENCE AND TECHNOLOGY 2025年第1期36卷

作者： Shao, Pengyan Lu, Tongwei Wuhan Inst Technol Sch Comp Sci & Engn Wuhan Hubei Peoples R China Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan Peoples R China

In the realm of industrial manufacturing, anomaly detection technologies are essential for enhancing production efficiency. This paper aims to optimize the detection of anomalies in industrial manufacturing settings by introducing a novel framework manifest and latent anomaly detection (MLAD). Firstly, we redefine anomalies as manifest and latent, further subdividing the original categories of structural and logical anomalies. To address the issue of insufficient feature space information utilization in previous reconstruction-based methods, which led to inadequate detection of latent anomalies, we combine a reconstruction-based approach with the multi-layer feature pyramid k-nearest neighbors technique. Additionally, we introduce a Hybrid Attention mechanism to refine the autoencoder's reconstruction process, thereby enhancing the detection of latent anomalies and increasing overall accuracy. Our experiments on the MVTec LOCO AD dataset demonstrate that MLAD achieves a significant improvement in anomaly detection, with an Image-level AUROC score of 90.8% and Pixel-level AU-sPRO of 79.3%. The integration of the Hybrid Attention mechanism significantly boosts the precision of anomaly detection, highlighting the effectiveness of our approach in enhancing industrial anomaly detection capabilities. The framework provides a more nuanced understanding of industrial defects, contributing valuable insights and practical improvements to the field of industrial anomaly detection.

关键词： industry anomaly detection anomaly classification MVTec LOCO AD dataset

来源：评论

学校读者我要写书评

暂无评论

Dual-stream network with complementary fusion and hierarchical attention for image tampering localization

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2025年第3期19卷 1-14页

作者： Mao, Zhanpeng Lu, Tongwei Wuhan Inst Technol Sch Comp Sci & Engn Wuhan 430205 Hubei Peoples R China Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan 430205 Hubei Peoples R China

In recent years, with the rapid development of image editing technology, the trustworthiness of multimedia data is facing severe challenges, and the security risks caused by image tampering are increasing, which promotes the research of image tampering localization. Traditional image tampering localization methods usually rely on identifying specific forgery traces, such as JPEG artifacts, edge inconsistencies, camera noise, and so on, and localize the tampered regions by cross-entropy loss. With the rise of deep learning, recent studies explore the boundary difference between tampered regions and real regions by adding Convolutional Neural Network and attention mechanism. However, there are two main weaknesses in these methods. First, the existing two-stream networks lack complementary fusion between blocks, and the features of the two branches lack interaction. The second is that they tend to focus on only a few specific forgery artifacts. In fact, the tampered images in real life may be generated by a variety of forgery methods, leaving a variety of tampering traces. To solve these problems, this paper proposes a novel image tampering Location network based on dual-stream dominant multi-step complementary fusion feature and multi-scale attention enhancement feature representation. The network combines the RGB feature and noise feature of the image to accurately identify and locate the possible tampering traces in the image. Specifically, our proposed RGB-Noise Complementary Fusion Module uses the weighting mechanism to complementary correct RGB features and Noise features, and fuses features to compensate for the lack of single-stream information while highlighting important features. In addition, Hierarchical Attention Module processes features at four scales at multiple levels, enabling the network to capture not only local details, but also global structural information, helping the model more accurately locate tampering regions. Experimental results show that

关键词： Image editing Tamper localization Dual-stream encoder-decoder network RGB-noise complementary fusion module Hierarchical attention module

来源：评论

学校读者我要写书评

暂无评论

intelligent classification framework for gear surface damage and gear type using CNN transfer learning

引用

PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART C-JOURNAL OF MECHANICAL ENGINEERING SCIENCE 2024年第9期238卷 4229-4239页

作者： Wang, Xiaopeng Hua, Hongpeng Peng, Kun Wu, Biqiong Xu, Xiang Du, Kanghua China Three Gorges Univ Yichang Key Lab Robot & Intelligent Yichang Peoples R China China Three Gorges Univ Hubei Key Lab Hydroelect Machinery Design & Mainte Yichang Peoples R China China Yangtze Power Co Ltd Hubei Key Lab Intelligent Yangtze & Hydroelect Sci Yichang Hubei Peoples R China Three Gorges Hydropower Plant China Yangtze Power Yichang Peoples R China ChinaThree Gorges Univ Yichang Key Lab Robot & Intelligent Yichang 443002 Hubei Peoples R China

In the mechanical transmission system gear has a very important role, and Gear surface damage is an important factor affecting gear transmission. Regular inspection of each gear and its surface damage is of great significance for ensuring the stable operation of a whole mechanical system. To improve the identification efficiency and accuracy for gears and their surface damage, In this paper, we propose a gear and its surface damage recognition method based on PyTorch deep learning library, expand the dataset using data augmentation and data extension techniques, compare the performance of gear and its surface damage recognition under three typical models, namely AlexNet, VGG16, and ResNet-101, and optimize each training model through migration learning and hyperparameter comparison. The results show that the speed and accuracy of model training can be improved by applying transfer learning. By extending the dataset using data augmentation techniques, the robustness of the network is improved considerably. When the batch size is 6 and the initial learning rate is set to 0.01, the model training effect is the best, and ResNet has higher recognition accuracy and stability than AlexNet and VGG16, which is more suitable for the classification of gears and their surface damage. In order to find the optimal ResNet model, four ResNet models with different number of layers, ResNet-34, ResNet-50, ResNet-101 and ResNet-152, were compared, among which the ResNet-101 model showed the optimal performance in gear surface damage recognition.

关键词： Convolutional neural network gear surface damage deep learning transfer learning

来源：评论

学校读者我要写书评

暂无评论

Exploring fusion domain: Advancing infrared and visible image fusion via IDFFN-GAN

引用

NEUROCOMPUTING 2025年 611卷

作者： Li, Juan Shi, Xiaoqian Li, Yanan Zhou, Huabing Wuhan Inst Technol Wuhan 430250 Peoples R China Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan 430250 Peoples R China

Infrared (IR) and visible (VI) images are crucial in applications such as surveillance and night vision, where each modality provides complementary information-IR captures thermal details, while VI captures textures. Fusing these images is essential to combine their strengths, resulting in a more comprehensive and informative image. In this work, we introduce the "fusion domain" concept, a unique distribution that optimally blends IR and VI features. Our primary contribution is developing the Intermediate Domain Feature Fusion Network (IDFFN), which employs an MLP to learn the optimal domain factor for this fusion. Integrated with the IDFFN-GAN's dual discriminators, our approach refines the fusion process to produce images that better preserve thermal and texture details from the proposed fusion and gradient domains. Experimental results demonstrate that our method achieves a 5% boost in Entropy (EN), a 50% rise in Spatial Frequency (SF), a 58% improvement in Average Gradient (AG), a 10% enhancement in Standard Deviation (SD), and a remarkable 198% increase in N AB / F compared to state-of-the-art techniques, ensuring superior retention of specific features from both modalities.

关键词： Image fusion Generative adversarial network Intermediate domain Domain factor Dual discriminators

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：