检索结果-内蒙古大学图书馆

polarization Guided HDR Reconstruction via pixel-Wise Depolarization

IEEE TRANSACTIONS ON IMAGE pROCESSING 2023年 32卷 1774-1787页

作者： Zhou, Chu Han, Yufei Teng, Minggui Han, Jin Li, Si Xu, Chao Shi, Boxin Peking Univ Sch Intelligence Sci & Technol Natl Key Lab Gen Artificial Intelligence Beijing 100080 Peoples R China Beijing Univ Posts & Telecommun Sch Artificial Intelligence Pattern Recognit & Intelligent Syst Lab Beijing 100876 Peoples R China Peking Univ Natl Engn Res Ctr Visual Technol Sch Comp Sci Beijing 100080 Peoples R China Univ Tokyo Grad Sch Informat Sci & Technol Tokyo 1138654 Japan

Taking photos with digital cameras often accompanies saturated pixels due to their limited dynamic range, and it is far too ill-posed to restore them. Capturing multiple low dynamic range images with bracketed exposures can make the problem less ill-posed, however, it is prone to ghosting artifacts caused by spatial misalignment among images. A polarization camera can capture four spatially-aligned and temporally-synchronized polarized images with different polarizer angles in a single shot, which can be used for ghost-free high dynamic range (HDR) reconstruction. However, real-world scenarios are still challenging since existing polarization-based HDR reconstruction methods treat all pixels in the same manner and only utilize the spatially-variant exposures of the polarized images (without fully exploiting the degree of polarization (Dop) and the angle of polarization (Aop) of the incoming light to the sensor, which encode abundant structural and contextual information of the scene) to handle the problem still in an ill-posed manner. In this paper, we propose a pixel-wise depolarization strategy to solve the polarization guided HDR reconstruction problem, by classifying the pixels based on their levels of ill-posedness in HDR reconstruction procedure and applying different solutions to different classes. To utilize the strategy with better generalization ability and higher robustness, we propose a network-physics-hybrid polarization-based HDR reconstruction pipeline along with a neural network tailored to it, fully exploiting the Dop and Aop. Experimental results show that our approach achieves state-of-the-art performance on both synthetic and real-world images.

关键词： High dynamic range imaging polarization deep learning

来源：评论

学校读者我要写书评

暂无评论

Urban Traffic pattern Analysis and Applications Based on Spatio-Temporal Non-Negative Matrix Factorization

引用

IEEE TRANSACTIONS ON intelligent TRANSpORTATION SYSTEMS 2022年第8期23卷 12752-12765页

作者： Wang, Yang Zhang, Yong Wang, Lixun Hu, Yongli Yin, Baocai Beijing Univ Technol Beijing Artificial Intelligence Inst Fac Informat Technol Beijing Key Lab Multimedia & Intelligent Software Beijing 100124 Peoples R China Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China Beijing Municipal Transportat Operat Coordinat Ct Beijing 100161 Peoples R China

Analyzing the traffic state of large citywide networks is an inherently difficult task. Various data issues, traffic signals, stops signs and other flow inhibitors of the network-level traffic state make the analysis more difficult than that under the small-scale local traffic state. To address this challenge, we propose a method based on spatio-temporal non-negative matrix factorization (ST-NMF), which is used for road network traffic pattern analysis. The method can be further extended to traffic data reconstruction and traffic prediction. In order to analyze traffic patterns, the proposed spatio-temporal non-negative matrix factorization model represents the network traffic as a linear combination of several basic patterns, which is also interpreted as the dynamics of spatial traffic characteristics over time in low-dimensional space. By the visual display of the spatial and temporal patterns and the assistance of clustering methods, the traffic pattern features are extracted. In the extended applications, data reconstruction relies on the sampling representation of missing data by ST-NMF, and data prediction is based on the prediction of the temporal patterns by ST-NMF. Through our method, we can not only obtain a high-quality data foundation, but also explore typical spatio-temporal patterns and general predictions of the future traffic state. The analysis results have important guiding significance on the management of intelligent transportation systems. Experiments on real-world traffic data are provided to verify the validity of our proposed approach.

关键词： Roads Transportation pattern analysis Data models Data mining Analytical models Autoregressive processes Traffic pattern analysis matrix factorization spatio-temporal characteristics

来源：评论

学校读者我要写书评

暂无评论

SiamON: Siamese Occlusion-Aware Network for Visual Tracking

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2023年第1期33卷 186-199页

作者： Fan, Chao Yu, Hongyuan Huang, Yan Shan, Caifeng Wang, Liang Li, Chenglong Anhui Univ Sch Artificial Intelligence Informat Mat & Intelligent Sensing Lab Anhui Prov Anhui Prov Key Lab Multimodal Cognit Computat Hefei 230601 Peoples R China Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 100190 Peoples R China Shandong Univ Sci & Technol Coll Elect Engn & Automat Qingdao 266590 Peoples R China Chinese Acad Sci Artificial Intelligence Res CAS AIR Beijing 100190 Peoples R China

Occlusion has been proven to be one of the most challenging factors faced by most visual trackers. There are mainly two difficulties, the first one is that the number of occlusion samples are very limited even though collecting a large-scale training data set, and another one is how to correctly learn the features of the target when comes to occlusion situations. In this paper, we tried to solve these two problems together in our proposed model. To this end, we propose a novel Siamese Occlusion-aware Network (SiamON) for high-performance visual tracking. In particular, we predefine some soft-masks to solve the problem of fewer occlusion samples, which perceive patterns of occlusion contents at different locations and take these masks as the conditions to guide occlusion-aware feature learning. Meanwhile, we propose a target-aware attention mechanism allows the model to pay more attention to the target and further weaken the impact of occlusion. Extensive experiments on several popular benchmarks show that our tracking method exceeds many state-of-the-art trackers especially in the presence of occlusion and meets the requirements of real-time.

关键词： Visual tracking occlusion aware attention siamese network

来源：评论

学校读者我要写书评

暂无评论

A Unified Framework for High Fidelity Face Swap and Expression Reenactment

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2022年第6期32卷 3673-3684页

作者： peng, Bo Fan, Hongxing Wang, Wei Dong, Jing Lyu, Siwei Chinese Acad Sci Inst Automat Ctr Res Intelligent Percept & Comp CRIPAC Natl Lab Pattern Recognit NLPR Beijing 100190 Peoples R China Chinese Acad Sci Inst Informat Engn State Key Lab Informat Secur Beijing 100093 Peoples R China Univ Buffalo State Univ New York Buffalo Sch Engn & Appl Sci Buffalo NY 14260 USA

Face manipulation techniques improve fast with the development of powerful image generation models. Two particular face manipulation methods, namely face swap and expression reenactment attract much attention for their flexibility and ease to generate high quality synthesis results. Recently, these two subjects are actively studied. However, most existing methods treat the two tasks separately, ignoring their underlying similarity. In this paper, we propose to tackle the two problems within a unified framework that achieves high quality synthesis results. The enabling component for our unified framework is the clean disentanglement of 3D pose, shape, and expression factors and then recombining them for different tasks accordingly. We then use the same set of 2D representations for face swap and expression reenactment tasks that are input to a common image translation model to directly generate the final synthetic images. Once trained, the proposed model can accomplish both face swap and expression reenactment tasks for previously unseen subjects. Comprehensive experiments and comparisons show that the proposed method achieves high fidelity results in multiple aspects, and it is especially good at faithfully preserving source facial shape in the face swap task, and accurately transferring facial movements in the expression reenactment task.

关键词： Faces Task analysis Videos Shape Three-dimensional displays Face recognition information integrity Face swap expression reenactment 3DMM video manipulation

来源：评论

学校读者我要写书评

暂无评论

Event-triggering load frequency control for multi-area power system based on random dynamic triggering mechanism and two-side closed functional?

引用

ISA TRANSACTIONS 2023年 133卷 193-204页

作者： Liu, Xingyue Shi, Kaibo Ma, Changyou Tang, Yiqian Tang, Lin Wei, Youhua Han, Yingjun Chengdu Univ Sch Elect Informat & Elect Engn Chengdu 610106 Peoples R China Anhui Univ Engn Res Ctr Power Qual Minist Educ Hefei 230601 Peoples R China Univ Elect Sci & Technol China Guangdong Inst Elect & Informat Engn Guangzhou 523808 Peoples R China Neijiang Normal Univ Coll Math & Informat Sci Data Recovery Key Lab Sichuan Prov Neijiang 641100 Peoples R China Chengdu Univ Technol Geomath Key Lab Sichuan Prov Chengdu 610059 Peoples R China China Tobacco Sichuan Ind Co LTD chengdu Cigarette Chengdu 610066 Peoples R China Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat Pr Chengdu 610106 Peoples R China

By taking into account sampled-data mechanism and transmission delay, the novel event-triggering load frequency control (LFC) strategy involving random dynamic triggering algorithm (RDTA) is developed for multi-area power systems in this paper. Firstly, an improved multi-area LFC model considering sampling and transmission delay (STD) simultaneously is addressed. Secondly, a modified event-triggering mechanism (ETM) with RDTA is proposed, considering parameter disturbances and a dynamic adjustment mechanism of the triggering threshold. Thirdly, a more advanced Lyapunov- Krasovskii functional (LKF) is constructed, introducing the delay-dependent matrices, more variable cross terms and the two-sided closed functional. Furthermore, two less conservative stability criteria are obtained according to the designed approach. Finally, two multi-area LFC systems are presented to verify the progressiveness of the proposed approach. (c) 2022 ISA. published by Elsevier Ltd. All rights reserved.

关键词： Event-triggering load frequency control Random dynamic triggering algorithm Two-side closed functional Multi-area load frequency control system Advanced Lyapunov-Krasovskii functional

来源：评论

学校读者我要写书评

暂无评论

On Degree Based Topological properties of Two Carbon Nanotubes

引用

pOLYCYCLIC AROMATIC COMpOUNDS 2022年第3期42卷 866-884页

作者： Zhang, Xiujun Rauf, Abdul Ishtiaq, Muhammad Siddiqui, Muhammad Kamran Muhammad, Mehwish Hussain Chengdu Univ Inst Higher Educ Sichuan Prov Key Lab Pattern Recognit & Intelligent Informat P Chengdu Peoples R China Air Univ Dept Comp Sci & Engn Multan Campus Multan Pakistan Comsats Univ Islamabad Dept Math Islamabad Pakistan Zhengzhou Univ Coll Chem Sch Chem Engn & Energy Zhengzhou Peoples R China

A nanostructure belong to a significant and an extensively investigated compounds in chemical science. It has been derived through engineering mechanism at the molecular scale. The most important of these new materials are carbon nanotubes. Carbon nanosheets are 2-dimensional lattices of carbon nanotubes. To compute and study topological indices of nanostructures are a respected problem in nanotechnology. Quantitative structure-property and structure-activity relationships of the carbon nanotubes compounds necessitate expressions for the molecular topological features of these compounds. Topological indices are vital devices for investigating chemical compounds to comprehend the fundamental topology of chemical structures. Ev-degree and ve-degree based topological indices are two novel degrees based indices as of late defined in chemical graph theory. In this paper, we have computed topological indices based on ev-degree and ve-degree for the two significant classes of carbon nanosheets composed by C-4 and C-8 denoted by and (TC4C8)-C-1 (S)[p,q] (TC4C8)-C-2(R)[p,q].

关键词： Carbon nanosheets ev-degree (TC4C8)-C-1 (S)[p,q] (TC4C8)-C-2(R)[p,q] topological indices ve-degree

来源：评论

学校读者我要写书评

暂无评论

The study on spatial distribution of water ecological environment carrying capacity during extreme drought conditions

引用

SCIENTIFIC REpORTS 2024年第1期14卷 1-13页

作者： Zhou, Yang Gui, Yingying Zhou, Qiang Li, Li Chen, Miaomiao Liu, Yuling Chengdu Univ Sch Architecture & Civil Engn Chengdu 610106 Sichuan Peoples R China Chengdu Univ Key Lab Pattern Recognit & Intelligent Informat Pr Chengdu 610106 Peoples R China Xian Univ Technol State Key Lab Ecohydraul Northwest Arid Reg Xian 710048 Shanxi Peoples R China Chengdu Univ Meat Proc Key Lab Sichuan Prov Chengdu Peoples R China Chengdu Univ Coll Comp Sci Chengdu 610106 Sichuan Peoples R China Sichuan Acad Agr Machinery Sci Chengdu Peoples R China Xian Univ Technol Xian 710048 Shanxi Peoples R China

Due to global warming and the disturbance of the interannual variability of precipitation, the frequency of extreme drought events has increased. The impact of global climate change on water resources is becoming increasingly apparent, then it is particularly necessary to explore the carrying capacity of water ecological environment under extreme drought conditions, which can guarantee the ecological water security in river basins. This study takes the Guanzhong area of the Wei River Basin as an example, calculating the water environment carrying capacity of 40 areas in the Weihe Guanzhong area in different levels of years under extreme drought conditions by comprehensive evaluation model of carrying capacity and using geographic information system GIS to display the spatial distribution of water environment carrying capacity in 40 regions. According to the results of the spatial distribution of water environmental bearing capacity, four different schemes are designed to improve the bearing capacity. The first plan reduces the industrial water consumption and irrigation quota by 5%, the second plan increases the industrial water and sewage treatment rate on this basis. the third plan further improves the development and utilization rate of surface and groundwater, and the fourth plan, on the basis of the first three plans, supplies 600 million cubic meters of industrial and agricultural water to Guanzhong region. Through comparative analysis, without taking any measures, under the extreme drought conditions, the water environment carrying capacity of the 40 areas in Guanzhong is all in an unbearable state. Overall, plan 4 has the most significant improvement in the water environment-carrying capacity, especially the Dong zhuang Reservoir of the Jing River which has played a very important role in enhancing the water ecological environment carrying capacity of the downstream water of the Wei River.

关键词： Water ecological environment-carrying capacity Wei River Water environment indicator

来源：评论

学校读者我要写书评

暂无评论

Tp>2p>EA: Target-Aware Taylor Expansion Approximation Network for Infrared and Visible Image Fusion

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2025年第5期35卷 4831-4845页

作者： Huang, Zhenghua Lin, Cheng Xu, Biyun Xia, Menghan Li, Qian Li, Yansheng Sang, Nong Wuhan Inst Technol Hubei Key Lab Opt Informat & Pattern Recognit Wuhan 430205 Peoples R China Wuchang Univ Technol Artificial Intelligence Sch Wuhan 430223 Peoples R China Harbin Engn Univ Coll Intelligent Syst Sci & Engn Harbin 150001 Peoples R China Wuhan Donghu Univ Sch Elect Informat Engn Wuhan 430212 Peoples R China Wuhan Univ Sch Remote Sensing & Informat Engn Wuhan 430079 Peoples R China Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Wuhan 430074 Peoples R China

In the image fusion mission, the crucial task is to generate high-quality images for highlighting the key objects while enhancing the scenes to be understood. To complete this task and provide a powerful interpretability as well as a strong generalization ability in producing enjoyable fusion results which are comfortable for vision tasks (such as objects detection and their segmentation), we present a novel interpretable decomposition scheme and develop a target-aware Taylor expansion approximation (T(2)EA) network for infrared and visible image fusion, where our T(2)EAincludes the following key procedures: Firstly, visible and infrared images are both decomposed into feature maps through a designed Taylor expansion approximation (TEA) network. Then, the Taylor feature maps are hierarchically fused by a dual-branch feature fusion (DBFF) network. Next, the fused map of each layer is contributed to synthesize an enjoyable fusion result by the inverse Taylor expansion. Finally, a segmentation network is jointed to refine the fusion network parameters which can promote the pleasing fusion results to be more suitable for segmenting the objects. To validate the effectiveness of our reported T(2)EAnetwork, we first discuss the selection of Taylor expansion layers and fusion strategies. Then, both quantitatively and qualitatively experimental results generated by the selected SOTA approaches on three datasets (MSRS, TNO, and LLVIp) are compared in testing, generalization, and target detection and segmentation, demonstrating that our T(2)EA can produce more competitive fusion results for vision tasks and is more powerful for image adaption.

关键词： Feature extraction Image fusion Taylor series Image segmentation Transformers Object detection Deep learning Correlation Transforms Object segmentation Taylor expansion approximation infrared and visible feature representation dual-branch feature fusion object detection

来源：评论

学校读者我要写书评

暂无评论

Learning pose-invariant 3D object reconstruction from single-view images

引用

NEUROCOMpUTING 2021年 423卷 407-418页

作者： peng, Bo Wang, Wei Dong, Jing Tan, Tieniu Chinese Acad Sci CASIA Ctr Res Intelligent Percept & Comp CRIPAC Natl Lab Pattern Recognit NLPR Inst Automat Beijing 100190 Peoples R China Guangdong Key Lab Intelligent Informat Proc Shenzhen 518060 Peoples R China Shenzhen Key Lab Media Secur Shenzhen 518060 Peoples R China

Learning to reconstruct 3D shapes using 2D images is an active research topic, with benefits of not requiring expensive 3D data. However, most work in this direction requires multi-view images for each object instance as training supervision, which oftentimes does not apply in practice. In this paper, we relax the common multi-view assumption and explore a more challenging yet more realistic setup of learning 3D shape from only single-view images. The major difficulty lies in insufficient constraints that can be provided by single view images, which leads to the problem of pose entanglement in learned shape space. As a result, reconstructed shapes vary along input pose and have poor accuracy. We address this problem by taking a novel domain adaptation perspective, and propose an effective adversarial domain confusion method to learn pose-disentangled compact shape space. Experiments on single-view reconstruction show effectiveness in solving pose entanglement, and the proposed method achieves on-par reconstruction accuracy with state-of-the-art with higher efficiency. (C) 2020 Elsevier B.V. All rights reserved.

关键词： Learning 3D shape Single view supervision Domain confusion Adversarial learning

来源：评论

学校读者我要写书评

暂无评论

Two-photon-pumped amplified spontaneous emission from Ruddlesden-popper perovskite flakes

引用

OpTICS EXpRESS 2022年第12期30卷 21094-21102页

作者： Li, Xiaohong Liu, Weiwei Gao, Yan Qin, Yan Long, Hua Wang, Kai Wang, Bing Lu, peixiang Wuhan Inst Technol Hubei Key Lab Opt Informat & Pattern Recognit Wuhan 430205 Peoples R China Huazhong Univ Sci & Technol Sch Phys Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Wuhan Natl Lab Optoelect Wuhan 430074 Peoples R China Guangdong Intelligent Robot Inst Dongguan 523808 Peoples R China

Herein, we report the two-photon pumped amplified spontaneous emission (ASE) in the 2D Rpps flakes at room temperature. We prepared high-quality (BA)(2)(MA)(n-1)pbnI3n+1 (n = 1, 2, 3, 4, 5) flakes by mechanical exfoliating from the fabricated crystals. We show that the (BA)(2)(MA)(n-1)pbnI3n+1 flakes display a tunable two-photon pumped emission from 527 nm to 680 nm, as n increases from 1 to 5. Furthermore, we demonstrated two-photon pumped ASE from the (BA)(2)(MA)(n-1)pbnI3n+1 (n = 3, 4, 5) flakes. The two-photon pumped ASE thresholds of the Rpps are lower than lots of the other semiconductor nanostructures, indicating an excellent performance of the Rpps for two-photon pumped emission. In addition, we investigated the pump-wavelength-dependent two-photon pumped ASE behaviors of the Rpps flakes, which suggest that the near-infrared laser in a wide wavelength range can be converted into visible light by the frequency upconversion process in Rpps. This work has opened new avenues for realizing nonlinearly pumped ASE based on the Rpps, which shows great potential for the applications in wavelength-tunable frequency upconversion. (C) 2022 Optica publishing Group under the terms of the Optica Open Access publishing Agreement

关键词： Amplified spontaneous emission Emission Laser arrays Second harmonic generation Spontaneous emission Visible light

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：