检索结果-内蒙古大学图书馆

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Liu, Hao Wang, Kun Han, Yudong Wang, Haocong Hu, Yupeng Wang, Chunxiao Nie, Liqiang School of Software Shandong University Jinan China Key Laboratory of Computing Power Network and Information Security Ministry of Education Qilu University of Technology Shandong Academy of Sciences Jinan China School of Computer Science and Technology Harbin Institute of Technology Shenzhen China

ISBN: (纸本)9798350368741

Masked Image Modeling (MIM), following "mask-and-reconstruct" scheme, is a promising self-supervised method to learn scalable visual representation. Studies indicate that selecting an effective mask strategy is vital for MIM. However, existing approaches often rely on static pre-defined priors, which limit their ability to adapt mask strategies dynamically for network optimization. In this paper, we focus on the learning process of the network and introduce human-like curriculum into MIM for dynamic representation refinement, and propose an end-to-end framework Curriculum Masked Image Modeling (CurMIM). CurMIM consists of two components: Mask Priority Measurer, which acts as a curriculum learner to determine mask priority values using the network's intrinsic state information, and Dual Adaptive Selector, which serves as a curriculum scheduler to create effective masks based on these values. With negligible extra parameters, our curriculum-based method consistently establishes noticeable improvements across varying model sizes and benchmarks, showing effectiveness and generalization. © 2025 IEEE.

关键词： Curriculum Learning Image Representation Masked Image Modeling Visual Pre-training

来源：评论

学校读者我要写书评

暂无评论

An In-depth Analysis of Spectroscopic Unmixing for Target Identification in Hyper Spectral Images 5th

An In-depth Analysis of Spectroscopic Unmixing for Target Id...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Yashaswini, K.A. Mishra, Awakash Zaidi, Taskeen Sumit School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India Karnataka Bangalore India Department of Computer Science and Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India

ISBN: (纸本)9789819780426

This summary offers an in-intensity evaluation of a system getting-to-know approach, particularly Spectroscopic Unmixing, for target identity in hyperspectral images. This parametric and pixel-smart linear decomposition technique generates spectral functions for each pixel, allowing the identity and extraction of various fabric signatures. The method is applied by fitting the information into a mathematical version, similarly used for unmixing. Different spectral fashions used within the literature have additionally been discussed. The research specializes in implementing the method and its applications for extracting items or substances in complex situations, including city, agricultural, or maritime imaging. Subsequently, the benefits and barriers of Spectroscopic Unmixing have been highlighted. The conclusion provides an insight into the performance of the technique for specific imaging eventualities. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

Analyzing Structural/Chemical Compositions of Combustion Pollutants Using Hyper Spectral Images 5th

Analyzing Structural/Chemical Compositions of Combustion Pol...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Singh, Yudhveer Poornima, Galiveeti Mishra, Awakash Nidhya, M.S. Department of Computer Science & Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India Karnataka Bangalore India

ISBN: (纸本)9789819780426

The latest advancements in air quality monitoring strategies have made it possible to appropriately discover the presence of air pollution along with particulates, gaseous compounds, and their chemicals. An extra comprehensive approach to quantifying air-pollutant stages is to install Hyper Spectral imagery of enclosed surroundings combined with superior sensors. Such a technique allows for the detection and quantification of pollutant debris's chemical and structural additives, which can be regularly too small for imaging through individual sensors. This paper studies Hyper Spectral Imaging to evaluate and identify combustion pollutants’ chemical and structural compositions. The consequences imply that this approach can offer dependable and accurate information, including the design of the contaminants, their size and vicinity, and a higher knowledge of emission assets. Moreover, the inexpensive method can detect and map pollutant levels in enclosed environments. It is a foundation for further investigation on using Hyper Spectral Imaging in air-friendly monitoring and analysis. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Air quality

来源：评论

学校读者我要写书评

暂无评论

Investigating the Applications of Sentimental Analysis Algorithms for Highly-Automated Delivery Applications 5th

Investigating the Applications of Sentimental Analysis Algor...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Bhuvana, J. Ali, Arif Pawaskar, Praveen Giridhar Savita Karnataka Bangalore India Department of Computer Science & Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India

ISBN: (纸本)9789819780426

Sentimental evaluation algorithms enable computer systems to stumble on, classify, and degree textual content sentiment. This era can be leveraged to evaluate consumer evaluations and provide remarks to corporations to be able to improve provider shipping. This observe pursuits to analyze the capability packages of those algorithms in pretty automatic shipping packages, inclusive of self sufficient drones and robots. Reading consumer sentiments using those algorithms can assist to stumble on client satisfaction or dissatisfaction in a timely fashion. This evaluation may be used to inform decision-making and to enhance carrier revel in for customers. The take a look at will hire a mixture of strategies, consisting of interviews, surveys, and quantitative analyses, to evaluate the effectiveness of this era in incredibly automatic shipping programs. The outcomes of this study will offer precious perception into the potential applications of sentimental evaluation algorithms for reinforcing customer support in computerized shipping systems.. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Wavelet-Based Time Series Forecasting for Wireless Sensor Networks 5th

Wavelet-Based Time Series Forecasting for Wireless Sensor Ne...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Murugan, R. Arya, Rakesh Kokila, S. Mishra, Awakash Karnataka Bangalore India Department of Computer Science and Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India

ISBN: (纸本)9789819780426

Wavelet-primarily based time series forecasting has been an effective device for Wi-Fi sensor network (WSN) packages. This approach has proven to provide correct forecasts for a selection of situations whilst enabling strength performance through decreased statistics storage and transmission bandwidth necessities. By means of extracting the vital time-related functions from sensor facts streams, records compression is executed at the same time as retaining the accuracy of the forecasts. Furthermore, with the aid of taking advantage of the multi resolution decomposition belongings of wavelets, this method can appropriately seize the temporal dynamics of the authentic signal. This permits better approximation of the underlying signal’s temporal correlations while as compared to other kinds of forecasting strategies. Additionally, wavelet-primarily based methods permit robust and green adaptation of the prediction fashions to changing sensor records, for this reason adapting the overall performance of the forecasts to exclusive situations. All these homes make wavelet-based time series forecasting an appealing choice for WSN programs. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Wireless sensor networks

来源：评论

学校读者我要写书评

暂无评论

Degenerate polarization entangled photon source based on a single Ti-diffusion lithium niobate waveguide in a polarization Sagnac interferometer

引用

Chinese Physics B 2023年第8期32卷 69-73页

作者：孙宇孙昌伟周唯杨然端家晨龚彦晓徐平祝世宁 National Laboratory of Solid State Microstructures and School of Physics Nanjing UniversityNanjing 210093China NanZhi Institude of Advanced Optoelectronic Integration Nanjing 210018China Institute for Quantum Information and State Key Laboratory of High Performance Computing College of Computer Science and TechnologyNational University of Defense TechnologyChangsha 410073China

Combining a Ti-diffusion periodically poled lithium niobate(PPLN)waveguide with a Sagnac interferometer,two opposite directions type-II spontaneous parametric down conversions(SPDC)occur coherently and yield a high brightness,high stability polarization entanglement *** source produces degenerate photon pairs at 1540.4 nm with a brightness of B=(1.36±0.03)×10^(6) pairs/(s·nm·m W).We perform quantum state tomography to reconstruct the density matrix of the output state and obtain a fidelity of F=0.983±*** high brightness and phase stability of our waveguide source enable a wide range of quantum information experiments operating at a low pump power as well as hold the advantage in mass production which can promote the practical applications of quantum technologies.

关键词： polarization entanglement source Sagnac interferometer quantum information lithium niobate waveguide

来源：评论

学校读者我要写书评

暂无评论

A Decision Tree-Based Approach to Object-Oriented Classification of Hyper Spectral Imagery 5th

A Decision Tree-Based Approach to Object-Oriented Classifica...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Mishra, Awakash Suneetha, K. Sumit Akshatha, Y. Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India Karnataka Bangalore India Department of Computer Science and Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India School of Computer Science and Engineering Presidency University Karnataka Bengaluru India

ISBN: (纸本)9789819780426

Hyperspectral imagery is a faraway sensing fact that gives one-of-a-kind spectral reflectance measurements across the electromagnetic spectrum. This form of imagery has various uses for monitoring the environment and tracking flora, wooded area cowl, and urban increase. However, successful analysis of hyperspectral imagery requires correct item-orientated classification, which may be challenging and time-consuming to acquire with traditional methods. This paper affords a choice tree-based method to the object-oriented type of hyperspectral imagery. This technique uses statistical class strategies and, blended with professional information, picks out the numerous objects described in the imagery. The proposed approach has been tested on hyperspectral imagery from exceptional environments, and the consequences reveal high accuracy and occasional processing times. This approach can offer a more accurate object-oriented class of hyperspectral imagery than conventional strategies. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Hyperspectral imaging

来源：评论

学校读者我要写书评

暂无评论

Characterizing the Weed Species of Agricultural Landscapes Using Hyper Spectral Image Analysis 5th

Characterizing the Weed Species of Agricultural Landscapes U...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Nagaraja, S.R. Mishra, Awakash Prabhu, A. Sivaraman, Himani School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India Karnataka Bangalore India Department of Computer Science and Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Manduwala Dehradun Navgaon India

ISBN: (纸本)9789819780426

This paper explores using hyper spectral photo evaluation to discover and characterize weed species inside agricultural landscapes. By using excessive-decision hyperspectral imaging, the specific spectral fingerprints of the different weed species attempting to penetrate the crop canopy can be diagnosed. This fact can then be compared to spectral libraries for the precise identity of the weed species gift. With these spectral facts, researchers can map the spread of weeds, understand the composition of weed species present, and verify weed responses to exceptional farming strategies or weather changes. This information can then assist in the development of extra effective and selective weed management strategies in agriculture. In conclusion, hyperspectral picture analysis may provide a treasured tool for characterizing and dealing with weed species within agricultural landscapes. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： characterizing hyperspectral weather

来源：评论

学校读者我要写书评

暂无评论

Dynamic-Feature-Based Object Tracking Using Real-Time Image Processing 5th

Dynamic-Feature-Based Object Tracking Using Real-Time Image ...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Mishra, Awakash Rengarajan Bhatt, Rahul Prince, P. Blessed Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India Karnataka Bangalore India Department of Computer Science & Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India School of Computer Science and Engineering Presidency University Karnataka Bengaluru India

ISBN: (纸本)9789819780426

This research explores the utility of today's real-time picture processing for dynamic-characteristic-primarily based object monitoring. Notably, this painting proposes a novel tracking method that combines an actual-time, place-primarily based convolutional neural community (R-CNN) with monitoring with the detection (TbD) technique. The proposed approach is for contemporary R-CNNs to constantly come across transferring objects in video frames and then modern-day the TbD technique to generate accurate item tracking information. The proposed system is evaluated on two benchmark datasets and compared to numerous tracking methods. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Exploring the Efficiency of Clustered Neural Network Aggregators for Time Series 5th

Exploring the Efficiency of Clustered Neural Network Aggrega...

引用

5th International Conference on Data science, Machine Learning and Applications, ICDSMLA 2023

作者： Sivaraman, Himani Tembhre, Sh obhit Mishra, Awakash Gobi, N. Department of Computer Science and Engineering School of Engineering and Computing Dev Bhoomi Uttarakhand University Navgaon Manduwala Dehradun India School of Computer Science and Engineering Presidency University Karnataka Bengaluru India Maharishi School of Engineering and Technology Maharishi University of Information Technology Uttar Pradesh Lucknow India Karnataka Bangalore India

ISBN: (纸本)9789819780426

This paper explores the performance of clustered neural community aggregators (CNNA) for time series. Through the years, time collection evaluation has allowed for insights into numerous phenomena, ranging from low cost to environmental and organic and clustered neural networks’ purpose to provide higher outcomes via mixing a couple of neural networks. However, their effectiveness has yet to be thoroughly studied in time collection. In this paper, several CNNA architectures are compared against their non-clustered equivalents, and the results are evaluated across an expansion of time collection datasets. Additionally, the overall performance of different clustering algorithms is studied in terms of clustering first-class and time cost. The outcomes show that, compared to non-clustered fashions, CNNA can substantially enhance the prediction performance of time collection and that the effectiveness of various types of clustering algorithms depends substantially on the kind of time collection analyzed. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： effectiveness fashions substantially

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：