检索结果-内蒙古大学图书馆

31st European Signal Processing Conference, EUSIPCO 2023

作者： Imamura, Kanami Nakamura, Tomohiko Takamune, Norihiro Yatabe, Kohei Saruwatari, Hiroshi Graduate School of Information Science and Technology The University of Tokyo Tokyo Japan Department of Electrical Engineering and Computer Science Tokyo University of Agriculture and Technology Tokyo Japan

ISBN: (纸本)9789464593600

In this paper, we propose algorithms for handling non-integer strides in sampling-frequency-independent (SFI) convolutional and transposed convolutional layers. The SFI layers have been developed for handling various sampling frequencies (SFs) by a single neural network. They are replaceable with their non-SFI counterparts and can be introduced into various network architectures. However, they could not handle some specific configurations when combined with non-SFI layers. For example, an SFI extension of Conv-TasNet, a standard audio source separation model, cannot handle some pairs of trained and target SFs because the strides of the SFI layers become non-integers. This problem cannot be solved by simple rounding or signal resampling, resulting in the significant performance degradation. To overcome this problem, we propose algorithms for handling non-integer strides by using windowed sinc interpolation. The proposed algorithms realize the continuous-time representations of features using the interpolation and enable us to sample instants with the desired stride. Experimental results on music source separation showed that the proposed algorithms outperformed the rounding- and signal-resampling-based methods at SFs lower than the trained SF. © 2023 European Signal Processing Conference, EUSIPCO. All rights reserved.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Bangla Speech Emotion Recognition Using 3D CNN Bi-LSTM Model 1st

Bangla Speech Emotion Recognition Using 3D CNN Bi-LSTM Model

引用

1st International Conference on Machine Intelligence and Emerging Technologies, MIET 2022

作者： Islam, Md. Riadul Akhand, M.A.H. Kamal, Md Abdus Samad Department of Computer Science and Engineering Khulna University of Engineering and Technology Khulna Bangladesh Graduate School of Science and Technology Gunma University Kiryu376-8515 Japan

ISBN: (纸本)9783031346187

Speech emotion recognition (SER) is the process of identifying the emotional characteristics of speech. The efficiency of emotional characteristics extracted from the speech is essential for SER performances. Neural networks based on various deep learning (DL) models have been investigated for SER. The existing SER studies mostly concentrate on a few languages with rich resources. SER for Bangla is very limited although it is a major mother language in the world. This study focused on Bangla SER (BSER) using 3D convolutional neural networks (CNN) and bidirectional long short-term memory networks (Bi-LSTM).The use of various speech signal transformations, using the Short-time Fourier Transform (STFT), Chroma STFT, and Mel-frequency Cepstral Coefficient (MFCC), is the significant challenge of this research. Transformed features of three different methodologies are integrated to set into the 3D CNN block for feature extraction. The time distributed layer’s input is set by flattening the deep features. To classify the emotion, the time distributed layer’s output is put into a bidirectional LSTM layer. The proposed model performed better than comparable recent methods when evaluated on a corpus of emotional speech in Bangla. © 2023, ICST Institute for computer sciences, Social Informatics and Telecommunications Engineering.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Evaluating mechanism of banana pseudo-stem retting using seawater:A cost-effective surface pre-treatment approach

引用

Journal of Bioresources and Bioproducts 2024年第3期9卷 322-335页

作者： Prince Hotor Ahmed H.Hassanin Osbert Akatwijuka Mohamed A.H.Gepreel Mitsuo Yamamoto Yukie Saito Ahmed Abdel-Mawgood Egypt-Japan University of Science and Technology Biotechnology ProgramAlexandria GovernorateNew Borg Al Arab 21934Egypt Alexandria University Department of Textile EngineeringAlexandria GovernorateBab Sharqi 5424041Egypt Egypt-Japan University of Science and Technology Material Science and EngineeringAlexandria GovernorateNew Borg Al Arab 21934Egypt Graduate School of Agricultural and Life Sciences The University of TokyoBunkyo-KuTokyo 113-8657Japan

Retting has been employed to extract natural fibers from agricultural wastes as a biological and cost-effective approach for *** its global abundance,banana pseudo-stem is a promising agro-waste for lignocellulosic fiber *** this study,fibers were extracted from the pseudo-stems after being pre-treated under four conditions using seawater at room temperature for up to 35 d Bacterial isolation from the fresh seawater sample and screening for ligninolytic ability were *** load as well as laccase and manganese peroxidase enzyme activity profile assay during the retting duration were *** transform infrared(FT-IR)and X-day diffraction(XRD)analyses were also examined for both pre-treated and untreated extracted *** results shows that six out of the eight bacterial isolates had the ability to degrade *** treatments(Raw stem+Raw seawater)and(Autoclaved stem+Raw seawater)recorded the highest viable bacterial load of 9.24×102 and 4.46×102 CFU,respectively,on the 14th day of the retting ***,the highest laccase and manganese peroxidase enzymes activity was recorded for(Raw stem+Raw seawater)and(Autoclaved stem+Raw seawater)treatments in the second to the third *** FT-IR spectra of the pre-treated fibers revealed relative reductions in peaks attributed to polysaccharides and other amorphous substances for all retting *** XRD diffractogram revealed that the crystallinity index(CI)of pre-treated fibers increased in all seawater retting treatment ***,the CI for fibers pre-treated under enzymatic conditions were enhanced even after five *** analysis for selected bacterial isolates showed homology to sequences of Bacillus velezensis,Shewanella sp.L8–5,and Citrobacter amalonaticus and Bacillus subtilis j8 *** these findings,it was suggested that physical,biological,and chemical actions were collectively involved in the seawater retting process of banana ps

关键词： Seawater retting Banana pseudo-stem Natural fiber Enzyme activity Colony forming unit

来源：评论

学校读者我要写书评

暂无评论

A Low Cost Triple-Node-Upset Self-Recoverable Latch Design for Nanoscale CMOS

引用

IEEE Transactions on Device and Materials Reliability 2025年第2期25卷 296-307页

作者： Bai, Na Zhang, Pengshuai Xu, Yaohua Guo, Yunhui Li, Hang Wang, Yi Wen, Xiaoqing Anhui University Information Materials and Intelligent Sensing Laboratory of Anhui Province Anhui Hefei230601 China National University of Defense Technology College of Computers Hunan Changsha410073 China Kyushu Institute of Technology Graduate School of Computer Science and Systems Engineering Fukuoka8208502 Japan

Due to the constant reduction of semiconductor feature sizes, the charge sharing effect is getting worse. The occurrence probability of triple-node upset (TNU), which has a significant impact on the circuit’s reliability, is rising in nanoscale digital circuits. This paper proposes a low overhead and TNU self-recovery latch (TNU-SR) based on the 28 nm CMOS technology to efficiently tolerate TNU and reduce the power dissipation and latency of the latch. The latch design proposed in this article adopts stacked transistors to reduce sensitive nodes and area overhead. The TNU-SR latch uses clock gating and high-speed transfer path technology, resulting in low power dissipation and low delay. For simulation results, compared with existing latches, it can be known that the TNU-SR latch performs well in terms of delay, power consumption, and area, so it is a good choice for radiation-hardened design for space applications. In addition, compared with the latest TNU self-recovery latch (HLTNURL), the TNU-SR latch proposed in this paper reduces the Power-Area-Delay Product (PADP) metric by 111.13%. © 2001-2011 IEEE.

关键词： CMOS integrated circuits

来源：评论

学校读者我要写书评

暂无评论

Survey of Online Exam Proctoring Model to Detect Cheating Behavior based on Face Recognition 8

Survey of Online Exam Proctoring Model to Detect Cheating Be...

引用

8th International Conference on Systems and Informatics, ICSAI 2022

作者： Rabiha, Sucianna Ghadati Kartowisastro, Iman Herwidiana Setiawan, Reina Budiharto, Widodo Bina Nusantara University Binus Graduate Program - Computer Science. Computer Science Department Jakarta11480 Indonesia Bina Nusantara University Faculty of Engineering Computer Science Department Binus Graduate Program - Computer Science Computer Engineering Department Jakarta11480 Indonesia Bina Nusantara University School of Computer Science Computer Science Department Binus Graduate Program - Computer Science Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9781665474481

Exams are an important component of any educational program, including online education. In any test, there is a possibility of cheating, so its detection and prevention is important. This study aims to conduct an in-depth study of the online exam monitoring model approach based on facial recognition used to detect cheating. Based on the inclusion and exclusion criteria designed, 13 selected studies were obtained. From these studies, we conducted further analysis regarding the Face Detection Method, Face Recognition Method, Initial Feature, Behavior Analysis and Evaluation Metrics used in each study so as to provide answers to research questions. the most frequently used Face detection method was Viola-Jones with a presentation of 20%, then CNN and MTCNN with a total presentation of 21%. The most widely used face recognition method in selected studies is CNN and metrics Accuracy is one of the most frequently used evaluations with a percentage of 33%. While the features that are usually used to detect cheating during online exams include facial motion and head pose which occupies the first position. The second is eye movement, then multiple faces gaze estimation and facial expression is in third place. Other features that also play a role in analyzing cheating behavior are mouth detection, facial vector, landmark location, gesture and posture. © 2022 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

P2P-assisted distributed-server content delivery of 8K media : a DLT-based approach

引用

Multimedia Tools and Applications 2024年 1-38页

作者： Barlas, Gerassimos El-Fakih, Khaled Hiromori, Akihito Yamaguchi, Hirozumi Department of Electrical and Computer Engineering American University in Dubai Dubai Dubai United Arab Emirates Computer Science & amp Engineering Dept American University of Sharjah Sharjah Sharjah26666 United Arab Emirates Graduate School of Information Science and Technology Osaka University Osaka Japan

In this paper we study the problem of multi-source, P2P-assisted, multimedia content delivery of 8K content. Such media constitute a big challenge for contemporary VoD services, as their media rate can surpass the clients’ upload/download speeds. Our proposed methods tackle this shortcoming while also allowing a reduction in the server resource consumption. Two methods are presented and evaluated, that include global and local multi-installment schemes, that are shown to surpass server-only or non-DLT-based content distribution by a fair margin, in all the significant metrics. Our methods can either offer a significant improvement in access time or a boost in the number of clients served, while at the same time reducing the consumed server resources. The performance of the proposed methods is evaluated with a rigorous simulation study based on the current state of the Internet on two geographic locations that represent best-case and average-case scenarios. Our findings suggest that the methodology used to assign the content parts to individual sources, has a big influence of the consumption of server resources, the system’s scalability and service quality. One of the proposed methods (LMI) achieves a maximum of 70.5% improvement on the average client access time, while saving up to 23% of server resources. Based on the system settings the reduced access-time/faster access to content can be balanced against an increase in the served clients by up to 25%. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Multimedia systems

来源：评论

学校读者我要写书评

暂无评论

Social Network Analysis and PowerBI Dashboard to Determine E-Wallet Marketing Strategies 4

Social Network Analysis and PowerBI Dashboard to Determine E...

引用

4th International Conference on Intelligent Cybernetics technology and Applications, ICICyTA 2024

作者： Rita, Rita Monica, Sheila Aurelia, Miracle Girsang, Abba Suganda Puspita, Virienia Binus Business School Bina Nusantara University Jakarta11480 Indonesia Binus Graduate Program Bina Nusantara University Computer Science Department Jakarta11480 Indonesia Binus Graduate Program Bina Nusantara University Communication Department Jakarta11480 Indonesia

ISBN: (纸本)9798331506490

The rapid growth of digital payment platforms like OVO, ShopeePay, and GoPay in Indonesia has driven the need for businesses to optimize marketing strategies by analyzing customer interactions through social media. This study leverages Social Network Analysis (SNA) and Power BI to extract insights from Twitter data, offering a comprehensive view of how these platforms engage with consumers and shape brand perceptions. By applying sentiment analysis and classification using IndoBERT results, this study aims to identify key influencers, assess the sentiment around these brands, and visualize marketing communication patterns. The integration of SNA and BI tools provides a detailed, data-driven approach to improving business decision-making in the competitive digital payment market, by utilizing them to evaluate e-wallet tweet activity and user interactions in Indonesia and identifying key periods and regions for targeted marketing. Strategies like loyalty programs and segmented campaigns are recommended to enhance user engagement, address security concerns, and maximize market reach during culturally significant periods such as Eid and the holiday season. © 2024 IEEE.

关键词： Business Intelligence Digital Wallet Marketing Strategy Social Media Social Network Analysis

来源：评论

学校读者我要写书评

暂无评论

HKDSME: Heterogeneous Knowledge Distillation for Semi-supervised Singing Melody Extraction Using Harmonic Supervision 24

HKDSME: Heterogeneous Knowledge Distillation for Semi-superv...

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Yu, Shuai He, Xiaoliang Chen, Ke Yu, Yi School of Computer Science and Technology Donghua University Shanghai China University of California San Diego San Diego United States Graduate School of Advanced Science and Engineering Hiroshima University Hiroshima Japan

ISBN: (纸本)9798400706868

Singing melody extraction is a key task in the field of music information retrieval (MIR). However, decades of research works have uncovered two difficult issues. First, binary classification on frequency-domain audio features (e.g., spectrogram) is regarded as the primary method, which ignores the potential associations of musical information at different frequency bins, as well as their varying significance for output decisions. Second, the existing semi-supervised singing melody extraction models ignore the accuracy of the generated pseudo labels by semi-supervised models, which largely limits the further improvements of the model. To solve the two issues, in this paper, we propose a heterogeneous knowledge distillation framework for semi-supervised singing melody extraction using harmonic supervision, termed as HKDSME. We begin by proposing a four-class classification paradigm for determining the results of singing melody extraction using harmonic supervision. This enables the model to capture more information regarding melodic relations in spectrograms. To improve the accuracy issue of pseudo labels, we then build a semi-supervised method by leveraging the extracted harmonics as a consistent regularization. Different from previous methods, it judges the availability of unlabeled data in terms of the inner positional relations of extracted harmonics. To further build a light-weight semi-supervised model, we propose a heterogeneous knowledge distillation (HKD) module, which enables the prior knowledge to transfer between heterogeneous models. We also propose a novel confidence guided loss, which incorporates with the proposed HKD module to reduce the wrong pseudo labels. We evaluate our proposed method using several well-known public available datasets, and the findings demonstrate the efficacy of our proposed method. © 2024 ACM.

关键词： harmonic supervision

来源：评论

学校读者我要写书评

暂无评论

Experimental study on open-end correction of a Helmholtz resonator narrow slit neck

Noise Control Engineering Journal

引用

Noise Control Engineering Journal 2022年第2期70卷 128-135页

作者： Sakamoto, Shuichi Akiba, Yuki Iizuka, Ryo Watanabe, Shunsuke Kominami, Ryota Shintani, Tetsushi Mechanical Engineering Program Faculty of Engineering Niigata University Niigata Japan Graduate School of Science and Technology Niigata University Niigata Japan

In general, the acoustic effective length of the open end of an air column is longer than its geometric length. Here, the effective increment of the geometric length of an air column is called the open-end correction length, which is an acoustically crucial issue. The purpose of this study is to experimentally clarify the open-end correction length for a slit shape with a narrow clearance. For this reason, the sound absorption coefficient was measured using a sample in which the neck of the Helmholtz resonator was shaped as a slit with a narrow clearance. Next, the open-end correction length (i.e., mass end correction length and resistive end correction length) was calculated backward by comparing the experimental and theoretical sound absorption curves. The resistive end correction increased as the slit width became narrower within the range of 0.5 mm or less for both the center slit and the slit near the wall, indicating a significant discrepancy between the theoretical equation and the estimated value. For the mass end correction, in the case the center slit, the estimated value is close to the theoretical value of in the slit width range from 0.3 mm to 1.0 mm. The results of this study may be useful in eliminating discrepancies in correction values for open-end correction in slits with narrow clearance. © 2022 Institute of Noise Control Engineering.

关键词： Acoustic wave absorption

来源：评论

学校读者我要写书评

暂无评论

The Influence of Digital Financial Literacy on Interest in Investing in the Capital Market of Millennial Generation Students 1

引用

4th International Conference on Inventive Computation and Information Technologies, ICICIT 2022

作者： Gaol, Ford Lumban Christopher Dermawan, Gerry Rafif, Muhammad Resnandya Rubianto, Rully Matsuo, Tokuro Doctor of Computer Science Binus Graduate Program Jakarta Indonesia School of Information System Bina Nusantara University Alam Sutera Indonesia Advanced Institute of Industrial Technology Tokyo Japan

ISBN: (数字)9789811974021

ISBN: (纸本)9789811974014

In today's environment, financial knowledge is vital and should be possessed by everyone. Numerous people are financially disadvantaged as a result of the COVID-19 pandemic. This serves as a reminder of the critical nature of financial investment in the future. However, many Indonesians, particularly students, are still unfamiliar with capital market investing. In reality, investing in the stock market can provide a variety of benefits, including financial security, capital appreciation, and passive income. Financial literacy concerns can result in a lack of financial awareness among pupils. Indeed, financial literacy is widely available online and can be accessed at any time and from any location. We employed a quantitative method with a non-experimental research design for this study. A questionnaire survey is used to obtain data. The goal of this study is to determine whether there is a correlation between student interest in capital market investment and their digital financial literacy. As a result, extensive action is required to increase student interest in the capital market, which is precisely what this study wants to do. According to our findings, students who have access to digital financial and investment information have increased their financial literacy and developed a greater interest in investing in the capital market. Individuals are also encouraged by social media trends to try their hand at investing in the stock market. Additionally, students prefer to invest in stocks and mutual funds over other investing options. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Investments

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：