检索结果-内蒙古大学图书馆

29th IEEE International Conference on Image Processing, ICIP 2022

作者： Xu, Yuxiao Huo, Yongkai Song, Yukai College of Computer Science & Software Engineering Shenzhen University China

ISBN: (数字)9781665496209

ISBN: (纸本)9781665496209

Panoramic video is considered to be an attractive video format since it provides the viewers with an immersive experience. However, only the viewers' focused region of a panoramic video, viewport, is shown on the screen. Numerous applications may benefit from the future viewport prediction, such as the adaptive viewport delivery. Two aspects may be considered for predicting the viewports. Firstly, the viewport trajectory within a time interval is highly correlated, hence the historical viewport trajectory should be utilized. Secondly, the panoramic video content within the viewport highly correlates with human visual attention. In this paper, we propose the Emotional Attention map aided panoramic Viewport Estimation (EAVE) model for predicting the future viewport trajectory. In this model, the historical viewport trajectory and the emotional attention map are jointly exploited by a Long Short-Term Memory (LSTM) network. Our experiments show that the proposed method outperforms several baseline methods in terms of prediction accuracy on a panoramic dataset containing viewing trajectory data from 50 viewers for watching 75 panoramic videos. © 2022 IEEE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Advanced Bearing Fault Diagnosis Using Cuckoo Optimization and KAN Algorithms 4

Advanced Bearing Fault Diagnosis Using Cuckoo Optimization a...

引用

4th International Conference on Electronic Information engineering and computer science, EIECS 2024

作者： Wang, Lingchao Ai, Qiang Yan, Hailong Hao, Ming Li, Xiaoxin School of Information Engineering Liaoning Institute of Science and Engineering Jinzhou China College of Computer Qinghai Normal University Xining China Software College Liaoning Technical University Huludao China

ISBN: (纸本)9798331531409

This study proposes a bearing fault diagnosis method that combines the Cuckoo Optimization Algorithm (COA) with the KAN algorithm. COA, as an intelligent optimization algorithm, is primarily used to find the optimal hyperparameters of the model, thereby enhancing the efficiency and robustness of the diagnosis model. The KAN algorithm is employed to build an intelligent classification model, achieving accurate recognition of multiple fault modes. Experimental results show that the proposed method maintains high diagnostic accuracy even with limited data, providing an effective solution for bearing fault diagnosis. © 2024 IEEE.

关键词： Fault detection

来源：评论

学校读者我要写书评

暂无评论

ML R-CNN: A Location-Aware Network for Tooth Instance Segmentation 23

ML R-CNN: A Location-Aware Network for Tooth Instance Segmen...

引用

15th International Conference on Digital Image Processing, ICDIP 2023

作者： Wang, Jiajun Suo, Ruiyu Zhou, Yao College of Software Engineering Sichuan University China College of Physics Sichuan University China College of Computer Science Sichuan University China

ISBN: (纸本)9798400708237

Dentists work with hundreds of intraoral photos a day to diagnose patients' dental health, which is very time-consuming. Recently, deep learning methods have shown its potential in automated tooth localization and segmentation, thus reducing the workload of dentists and improving the diagnostic efficiency. However, teeth in different locations are similar in shape and size, which remains a serious challenge for obtaining accurate location and segmentation results. To cope with this issue, we propose a location-aware neural network, Mask Location R-CNN(ML R-CNN), which adopts a location head branch to combine location information with dental classification confidence. Besides, an improved backbone is introduced for enhancing the feature extraction. An intraoral photo dataset was constructed and annotated by senior dentists, and the experimental results show that the proposed method considerably improves performance of tooth instance segmentation comparing with other competitive approaches. © 2023 ACM.

关键词： Location

来源：评论

学校读者我要写书评

暂无评论

EvoSense: Towards Self-evolving WiFi-based User Gait Recognition 15

EvoSense: Towards Self-evolving WiFi-based User Gait Recogni...

引用

2022 IEEE Congress on Cybermatics: 15th IEEE International Conferences on Internet of Things, iThings 2022, 18th IEEE International Conferences on Green Computing and Communications, GreenCom 2022, 2022 IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2022 and 8th IEEE International Conference on Smart Data, SmartData 2022

作者： Yao, Yao Luo, Chengwen Feng, Xingyu Huang, Yijing Zhang, Jin Li, Jianqiang Shenzhen University College of Computer Science and Software Engineering China

ISBN: (数字)9781665454179

ISBN: (纸本)9781665454179

Motivated by the recent development in WiFi-based gait recognition, in this work we propose a new system named as EvoSense, which can dynamically adapt to new users and update the recognition models to achieve a robust recognition performance even when there are new/unknown users in the system. EvoSense adopts deep metric learning based on approaches to design an efficient feature extraction network, which is able to extract robust features for different users in WiFi gait recognition, and this makes it possible to dynamically cluster and detect new users and to update the recognition network to adapt to these new user automatically. This feature makes it appealing since WiFi gait recognition with new/unknown users is still an important open problem that hinders the wide deployment of such systems currently. We implement and evaluate the system using real-world dataset, and the results show that EvoSense achieves an accuracy of 93.3% on the dataset of 15 users with 5 additional unknown users, which shows the effectiveness of EvoSense in real-world applications. © 2022 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

HSI-DETR: A DETR-based Transfer Learning from RGB to Hyperspectral Images for Object Detection of Live and Dead Cells: To achieve better results, convert models with the fewest changes from RGB to HSI 11

HSI-DETR: A DETR-based Transfer Learning from RGB to Hypersp...

引用

11th International Conference on Computing and Pattern Recognition, ICCPR 2022

作者： Ye, Songxin Li, Nanying Xue, Jiaqi Long, Yaqian Jia, Sen College of Computer Science and Software Engineering Shenzhen University China

ISBN: (纸本)9781450397056

Traditional cell viability judgment methods are invasive and damaging to cells. Moreover, even under a microscope, it is difficult to distinguish live cells from dead cells by the naked eye alone. With the development of optical imaging technology, hyperspectral imaging is more and more widely used in various fields. Hyperspectral imaging is a non-contact optical technique that provides both spectral and spatial information in a single measurement. It becomes a fast, non-invasive option to differentiate between live and dead cells. In recent years, the rapid development of deep learning has provided a better way to distinguish the difference between living and dead cells through a large amount of data. However, it is often necessary to acquire large amounts of labeled data at an expensive cost to train models. This is more difficult to achieve on medical hyperspectral images. Therefore, in this paper, a new model called HSI-DETR is proposed to solve the above problem on the target detection task of live and dead cells, which is based on the detection transformer (DETR) model. The HSI-DETR model suitable for hyperspectral images (HSI) is proposed with minimal modification. Then, some parameters of DETR trained on RGB images are transferred to HSI-DETR trained on hyperspectral images. Compared to the general method, this method can train a better model with a small number of labeled samples. And compared to the DETR-R50, the AP50 of HSI-DETR-R50 has increased by 5.15%. © 2022 ACM.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Local Weighted Neighborhood Rough Sets for Weakly Labeled Data

Local Weighted Neighborhood Rough Sets for Weakly Labeled Da...

引用

2023 International Conference on Machine Learning and Cybernetics, ICMLC 2023

作者： Guo, Yanting Hu, Meng Liu, Qiang Tsang, Eric C.C. Wang, Xizhao College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China Macau University of Science and Technology Faculty of Innovation Engineering China

ISBN: (纸本)9798350303780

Local rough set models are efficient rough data analysis methods for weakly labeled data in the rough set field. However, the existing local models do not distinguish the importance of different attributes to decisions when describing and defining binary relations, resulting in inaccurate information granules or classes, which affects the accuracy of concept approximation. In this paper, we choose local neighborhood rough sets as the representative model, and then propose Local Weighted Neighborhood Rough Sets (LWNRS). First, the importance of attributes to decisions is quantified by combining the supervised information of labeled data and the structural information of unlabeled data, and an attribute weighted method based on semi-supervised dynamic optimization is proposed. Meanwhile, the weighted neighborhood relation and neighborhood classes of objects are defined under weakly labeled data, and the WLNRS model is constructed. In addition, the weighted dependency is proposed to measure the data uncertainty, and the relationships among WLNRS and other relevant local and global models are deeply explored. Finally, the efficiency and accuracy of WLNRS in concept approximation are verified by comparison. © 2023 IEEE.

关键词： Rough set theory

来源：评论

学校读者我要写书评

暂无评论

A Center-Masked Convolutional Transformer for Hyperspectral Image Classification 2

A Center-Masked Convolutional Transformer for Hyperspectral ...

引用

2nd Workshop on Complex Data Challenges in Earth Observation, CDCEO 2022

作者： Wang, Yifan Jiang, Shuguo Xu, Meng Zhang, Shuyu Jia, Sen College of Computer Science and Software Engineering Shenzhen University China

Hyperspectral images (HSIs) have a wide field of view and rich spectral information, where each pixel represents a small area of the earth's surface. The pixel-level classification task of HSI has become one of the research hotspots in hyperspectral image processing and analysis. More and more deep learning methods have been proposed in recent years, among which convolutional neural network (CNN) is the most influential. However, it is difficult for CNN-based models to obtain the global receptive field in HSI classification task. Besides, most of the self-supervised training methods are based on sample reconstruction, and it is not easy to achieve effective use of unlabeled samples. In this paper, we propose a novel convolutional embedding module, combined with the Transformer blocks, which successfully improves the context-awareness while retaining the local feature extraction capability. Moreover, a new self-supervised task is designed to make more efficient use of unlabeled data. Our proposed pre-training task only masks the central token and reconstructs the central pixel from a learnable vector. It allows the model to capture the patterns between the central object and surrounding objects without labels. © 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0)

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

Progressive Semantic Enhancement Network for Hyperspectral and LiDAR Classification

引用

IEEE Transactions on Neural Networks and Learning Systems 2024年第6期36卷 11327-11339页

作者： Xiyou Fu Xi Zhou Yawen Fu Pan Liu Sen Jia College of Computer Science and Software Engineering Shenzhen University Shenzhen China

The joint classification of hyperspectral image (HSI) and light detection and ranging (LiDAR) data is gaining attention for its improved classification accuracy. However, effectively integrating the rich spectral information of HSI and the elevation features of LiDAR has remained a challenge in multimodal fusion. This article proposes a novel approach called progressive semantic enhancement network (PSENet) for hyperspectral and LiDAR classification based on a progressive joint spatial-spectral attention mechanism. PSENet mainly comprises two modules: the spatial grouping constraint (SAGC) module and the spectral weighting constraint (SEWC) module. The SAGC module extracts multiscale features in the spatial domain, while the SEWC module focuses on enhancing semantic features in spectral dimension. By gradually utilizing spatial and spectral constraint modules to progressively enhance feature extraction, PSENet integrates affluent information for a more refined classification of ground objects. Based on experimental results, it has been demonstrated that PSENet outperforms several most advanced methods on three datasets. The SAGC and SEWC modules proposed in PSENet enable the effective integration of the spatial, spectral, and elevation information from HSI and LiDAR, providing a promising way to perform classification more accurately. The source codes of this work will be publicly available at http://***/ .

关键词： Laser radar Feature extraction Hyperspectral imaging Data mining Semantics Attention mechanisms Accuracy Distance measurement Neural networks Data models

来源：评论

学校读者我要写书评

暂无评论

Multiscale Spatial-spectral Joint Feature Learning for Multispectral and Hyperspectral Image Fusion 23

Multiscale Spatial-spectral Joint Feature Learning for Multi...

引用

23rd IEEE International Conference on High Performance Computing and Communications, 7th IEEE International Conference on Data science and Systems, 19th IEEE International Conference on Smart City and 7th IEEE International Conference on Dependability in Sensor, Cloud and Big Data Systems and Applications, HPCC-DSS-SmartCity-DependSys 2021

作者： Min, Zhichao Wang, Yifan Jia, Sen College of Computer Science and Software Engineering Shenzhen University China

ISBN: (纸本)9781665494571

For RGB image super-resolution, usually operates on a single image. However, due to a large number of spectral bands and high dimensionality of the data in hyperspectral images (HSIs), it is difficult for a single image super-resolution to improve spatial resolution without spectral distortion. Therefore, it is necessary to fuse spatial features of multispectral images (MSIs) to improve spatial resolution. In this paper, we take advantage of the end-to-end advantages of deep learning and propose multiscale spatial-spectral joint feature learning (MSSJFL) for multispectral and hyperspectral image fusion. MSSJFL method utilizes different scale feature learning to exact single spatial feature, spectral feature, spatial-spectral joint feature, and then restore spectral information and spatial details via feature restoration module. Experimental results demonstrate that the proposed network architecture achieves better performance compared with state-of-the-art HSI and MSI fusion. The classification accuracy of the fusion result on the real HSI dataset is also closest to the ground truth. The code of MSSJFL is available at https://***/jx-mzc/MSSJFL. © 2021 IEEE.

关键词： Representation learning Deep learning Smart cities Superresolution Network architecture Feature extraction Image restoration

来源：评论

学校读者我要写书评

暂无评论

A novel observation points‐based positive‐unlabeled learning algorithm

引用

CAAI Transactions on Intelligence Technology 2023年第4期8卷 1425-1443页

作者： Yulin He Xu Li Manjing Zhang Philippe Fournier‐Viger Joshua Zhexue Huang Salman Salloum Guangdong Laboratory of Artificial Intelligence and Digital Economy(SZ) ShenzhenChina College of Computer Science and Software Engineering Shenzhen UniversityShenzhenChina School of Computing National University of SingaporeSingaporeSingapore

In this study,an observation points‐based positive‐unlabeled learning algorithm(hence called OP‐PUL)is proposed to deal with positive‐unlabeled learning(PUL)tasks by judiciously assigning highly credible labels to unlabeled *** proposed OP‐PUL algorithm has three ***,an observation point classifier ensemble(OPCE)algorithm is constructed to divide unlabeled samples into two categories,which are temporary positive and permanent negative ***,a temporary OPC(TOPC)is trained based on the combination of original positive samples and permanent negative samples and then the permanent positive samples that are correctly classified with TOPC are retained from the temporary positive ***,a permanent OPC(POPC)is finally trained based on the combination of original positive samples,permanent positive samples and permanent negative *** exhaustive experimental evaluation is conducted to validate the feasibility,rationality and effectiveness of the OP‐PUL algorithm,using 30 benchmark PU data *** show that(1)the OP‐PUL algorithm is stable and robust as unlabeled samples and positive samples are increased in unlabeled data sets and(2)the permanent positive samples have a consistent probability distribution with the original positive ***,a statistical analysis reveals that POPC in the OP‐PUL algorithm can yield better PUL performances on the 30 data sets in comparison with four well‐known PUL *** demonstrates that OP‐PUL is a viable algorithm to deal with PUL tasks.

关键词： artificial intelligence datamining,machine learning

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：