检索结果-南通市图书馆

维普期刊数据库

No Reference Image Sharpness Assessment Based on Global Color Difference Variation

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2024年第1期33卷 293-302页

作者： Chenyang SHI Yandan LIN School of artificial intelligence Anhui Polytechnic University School of Mechanical Engineering Anhui Polytechnic University Department of Light Sources and Illuminating Engineering Fudan University Academy for Engineering & Technology Fudan University

Image quality assessment(IQA) model is designed to measure the image quality in consistent with subjective ratings by computational models. In this research, a valid no reference IQA(NR-IQA) model for color image sharpness assessment is proposed based on local color difference map in a color space. In the proposed model, the absolute color difference variation and relative color difference variation are combined to evaluate sharpness in YIQ color space(a color coordinate system for the development of the United States color television system). The difference between sharpest and blurriest spot of an image is represented by the absolute color difference variation, and relative color difference variation expresses the variation in the image content. Extensive experiments are performed on five publicly availab.e benchmark synthetic blur databases and two real blur databases, and the results prove that the proposed model work better than the other state-of-the-art and latest NR-IQA models for the prediction accuracy on blurry images. Besides, the model maintains the lowest computational complexity.

关键词： Image quality TV Image color analysis Databases Computational modeling Color Predictive models

Extraction of intersecting palm-vein and palmprint features for cancellab.e identity verification

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

CAAI Transactions on intelligence Technology 2024年第1期9卷 69-86页

作者： Jaekwon Lee Beom-Seok Oh Kar-Ann Toh School of Electrical and Electronic Engineering Yonsei UniversitySeoulRepublic of Korea Department of Applied artificial intelligence Seoul National University of Science and TechnologySeoulRepublic of Korea

A novel method based on the cross-modality intersecting features of the palm-vein and the palmprint is proposed for identity *** on the unique geometrical relationship between the two biometric modalities,the cross-modality intersecting points provides a stable set of features for identity *** facilitate flexibility in template changes,a template transformation is *** maintaining non-invertibility,the template transformation allows transformation sizes beyond that offered by the con-ventional *** experiments using three public palm databases are conducted to verify the effectiveness the proposed system for identity recognition.

关键词： biometrics hand vein recognition hand-based biometrics multi-biometrics palmprint recognition

CAW-YOLO:Cross-Layer Fusion and Weighted Receptive Field-Based YOLO for Small Object Detection in Remote Sensing

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computer Modeling in Engineering & Sciences 2024年第6期139卷 3209-3231页

作者： Weiya Shi Shaowen Zhang Shiqiang Zhang College of artificial intelligence and Big Data Henan University of TechnologyZhengzhou450001China College of Information Science and Engineering Henan University of TechnologyZhengzhou450001China

In recent years,there has been extensive research on object detection methods applied to optical remote sensing images utilizing convolutional neural *** these efforts,the detection of small objects in remote sensing remains a formidable *** deep network structure will bring about the loss of object features,resulting in the loss of object features and the near elimination of some subtle features associated with small objects in deep ***,the features of small objects are susceptible to interference from background features contained within the image,leading to a decline in detection ***,the sensitivity of small objects to the bounding box perturbation further increases the detection *** this paper,we introduce a novel approach,Cross-Layer Fusion and Weighted Receptive Field-based YOLO(CAW-YOLO),specifically designed for small object detection in remote *** address feature loss in deep layers,we have devised a cross-layer attention fusion *** noise is effectively filtered through the incorporation of Bi-Level Routing Attention(BRA).To enhance the model’s capacity to perceive multi-scale objects,particularly small-scale objects,we introduce a weightedmulti-receptive field atrous spatial pyramid ***,wemitigate the sensitivity arising from bounding box perturbation by incorporating the joint Normalized Wasserstein Distance(NWD)and Efficient Intersection over Union(EIoU)*** efficacy of the proposedmodel in detecting small objects in remote sensing has been validated through experiments conducted on three publicly availab.e *** experimental results unequivocally demonstrate the model’s pronounced advantages in small object detection for remote sensing,surpassing the performance of current mainstream models.

关键词： Small object detection attention mechanism cross-layer fusion discrete cosine transform

Ensemble successor representations for task generalization in offline-to-online reinforcement learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第7期67卷 240-255页

作者： Changhong WANG Xudong YU Chenjia BAI Qiaosheng ZHANG Zhen WANG Space Control and Inertial Technology Research Center Harbin Institute of Technology Shanghai artificial intelligence lab.ratory Shenzhen Research Institute of Northwestern Polytechnical University School of Cybersecurity Northwestern Polytechnical University

In reinforcement learning(RL), training a policy from scratch with online experiences can be inefficient because of the difficulties in exploration. Recently, offline RL provides a promising solution by giving an initialized offline policy, which can be refined through online interactions. However, existing approaches primarily perform offline and online learning in the same task, without considering the task generalization problem in offline-to-online adaptation. In real-world applications, it is common that we only have an offline dataset from a specific task while aiming for fast online-adaptation for several tasks. To address this problem, our work builds upon the investigation of successor representations for task generalization in online RL and extends the framework to incorporate offline-to-online learning. We demonstrate that the conventional paradigm using successor features cannot effectively utilize offline data and improve the performance for the new task by online fine-tuning. To mitigate this, we introduce a novel methodology that leverages offline data to acquire an ensemble of successor representations and subsequently constructs ensemble Q functions. This approach enables robust representation learning from datasets with different coverage and facilitates fast adaption of Q functions towards new tasks during the online fine-tuning *** empirical evaluations provide compelling evidence showcasing the superior performance of our method in generalizing to diverse or even unseen tasks.

关键词： offline reinforcement learning online fine-tuning task generalization successor representations ensembles

同方期刊数据库

Proposing a machine learning approach to analyze and predict basic high-temperature properties of iron ore fines and its factors

在线全文

学校读者我要写书评

暂无评论

Journal of Iron and Steel Research(International) 2024年第5期31卷 1082-1094页

作者： Qing-ke Sun Yao-zu Wang Jian-liang Zhang Zheng-jian Liu Le-le Niu Chang-dong Shan Yun-fei Ma Institute of artificial intelligence University of Science and Technology BeijingBeijing100083China School of intelligence Science and Technology University of Science and Technology BeijingBeijing100083China School of Metallurgical and Ecological Engineering University of Science and Technology BeijingBeijing100083China

The basic high-temperature properties of iron ore play a crucial role in optimizing sintering and ore blending,but the testing process for these properties is complex and has significant lag time,which cannot meet the actual needs of ore blending.A prediction model for the basic high-temperature properties of iron ore fines was thus proposed based on a combination of machine learning algorithms and genetic ***,the prediction accuracy of different machine learning models for the basic high-temperature properties of iron ore fines was ***,a random forest model optimized by genetic algorithms was built,further improving the prediction accuracy of the *** test results show that the random forest model optimized by genetic algorithms has the highest prediction accuracy for the lowest assimilation temperature and liquid phase fluidity of iron ore,with a determination coefficient of 0.903 for the lowest assimilation temperature and 0.927 for the liquid phase fluidity after *** trained model meets the fluctuation requirements of on-site testing and has been successfully applied to actual production on site.

关键词： Iron ore Basic high-temperature property Machine learning Random forest Genetic algorithm

Novel Rifle Number Recognition Based on Improved YOLO in Military Environment

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第1期78卷 249-263页

作者： Hyun Kwon Sanghyun Lee Department of artificial intelligence and Data Science Korea Military AcademySeoulKorea Graduate School of Information Security Korea Advanced Institute of Science and TechnologyDaejeonKorea

Deep neural networks perform well in image recognition,object recognition,pattern analysis,and speech *** military applications,deep neural networks can detect equipment and recognize *** military equipment,it is necessary to detect and recognize rifle management,which is an important piece of equipment,using deep neural *** have been no previous studies on the detection of real rifle numbers using real rifle image *** this study,we propose a method for detecting and recognizing rifle numbers when rifle image data are *** proposed method was designed to improve the recognition rate of a specific dataset using data fusion and transfer *** the proposed method,real rifle images and existing digit images are fusedas trainingdata,andthe final layer is transferredto theYolov5 *** detectionand recognition performance of rifle numbers was improved and analyzed using rifle image and numerical *** used actual rifle image data(K-2 rifle)and numeric image datasets,as an experimental *** was used as the machine learning *** results show that the proposed method maintains 84.42% accuracy,73.54% precision,81.81% recall,and 77.46% F1-score in detecting and recognizing rifle *** proposed method is effective in detecting rifle numbers.

关键词： Machine learning deep neural network rifle number recognition detection

BMLP:behavior-aware MLP for heterogeneous sequential recommendation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Frontiers of Computer Science 2024年第3期18卷 235-237页

作者： Weixin LI Yuhao WU Yang LIU Weike PAN Zhong MING College of Computer Science and Software Engineering Shenzhen UniversityShenzhen 518060China Guangdong lab.ratory of artificial intelligence and Digital Economy(SZ) Shenzhen 518123China

1 Introduction Recommender systems can effectively alleviate the problem of information ***,traditional recommendation methods cannot capture users’dynamic *** recommendation methods model user sequences to obtain more accurate and dynamic user ***,deep learning-based sequential recommendation methods have achieved great *** is proposed to capture the sequential information[1,2].Attention-based methods[3]use attention mechanisms to learn relationships between ***-based methods[4−6]transform sequences into graph structures to capture relationships of ***,they have the following two limitations.

关键词： sequences behavior sequential

Perpendicular-Cutdepth:Perpendicular Direction Depth Cutting Data Augmentation Method

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第4期79卷 927-941页

作者： Le Zou Linsong Hu Yifan Wang Zhize Wu Xiaofeng Wang Anhui Provincial Engineering lab.ratory of Big Data Technology Application forUrban Infrastructure School ofArtificial Intelligence and Big DataHefei UniversityHefei230601China Institute of Applied Optimization School of Artificial Intelligence and Big DataHefei UniversityHefei230601China

Depth estimation is an important task in computer *** data at scale for monocular depth estimation is challenging,as this task requires simultaneously capturing RGB images and depth ***,data augmentation is crucial for this *** data augmentationmethods often employ pixel-wise transformations,whichmay inadvertently disrupt edge *** this paper,we propose a data augmentationmethod formonocular depth estimation,which we refer to as the Perpendicular-Cutdepth *** method involves cutting realworld depth maps along perpendicular directions and pasting them onto input images,thereby diversifying the data without compromising edge *** validate the effectiveness of the algorithm,we compared it with existing convolutional neural network(CNN)against the current mainstream data augmentation ***,to verify the algorithm’s applicability to Transformer networks,we designed an encoder-decoder network structure based on Transformer to assess the generalization of our proposed *** results demonstrate that,in the field of monocular depth estimation,our proposed method,Perpendicular-Cutdepth,outperforms traditional data *** the indoor dataset NYU,our method increases accuracy from0.900 to 0.907 and reduces the error rate from0.357 to *** the outdoor dataset KITTI,our method improves accuracy from 0.9638 to 0.9642 and decreases the error rate from 0.060 to 0.0598.

关键词： Perpendicular depth estimation data augmentation

Understanding adversarial attacks on observations in deep reinforcement learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2024年第5期67卷 69-83页

作者： You QIAOBEN Chengyang YING Xinning ZHOU Hang SU Jun ZHU Bo ZHANG Department of Computer Science and Technology Beijing National Research Center for Information Science and Technology Tsinghua-Bosch Joint Center for Machine Learning Institute for Artificial Intelligence Tsinghua University Peng Cheng lab.ratory

Deep reinforcement learning models are vulnerable to adversarial attacks that can decrease the cumulative expected reward of a victim by manipulating its observations. Despite the efficiency of previous optimization-based methods for generating adversarial noise in supervised learning, such methods might not achieve the lowest cumulative reward since they do not generally explore the environmental ***, a framework is provided to better understand the existing methods by reformulating the problem of adversarial attacks on reinforcement learning in the function space. The reformulation approach adopted herein generates an optimal adversary in the function space of targeted attacks, repelling them via a generic two-stage framework. In the first stage, a deceptive policy is trained by hacking the environment and discovering a set of trajectories routing to the lowest reward or the worst-case performance. Next, the adversary misleads the victim to imitate the deceptive policy by perturbing the observations. Compared to existing approaches, it is theoretically shown that our adversary is strong under an appropriate noise level. Extensive experiments demonstrate the superiority of the proposed method in terms of efficiency and effectiveness, achieving state-of-the-art performance in both Atari and MuJoCo environments.

关键词： deep learning reinforcement learning adversarial robustness adversarial attack