检索结果-内蒙古大学图书馆

Fine-Grained Emotion Prediction for Movie and Television scene images

The Journal of China Universities of Posts and Telecommunications 2024年第3期31卷 43-55页

作者： Su Zhibin Zhou Xuanye Liu Bing Ren Hui State Key Laboratory of Media Convergence and Communication Communication University of ChinaBeijing 100024China Key Laboratory of Acoustic Visual Technology and Intelligent Control System Communication University of ChinaBeijing 100024China School of Information and Communication Engineering Communication University of ChinaBeijing 100024China

For the task of content retrieval,analysis and generation of film and television scene images in the field of intelligent editing,fine-grained emotion recognition and prediction of images is of great *** this paper,the fusion of traditional perceptual features,art features and multi-channel deep learning features are used to reflect the emotion expression of different levels of the *** addition,the integrated learning model with stacking architecture based on linear regression coefficient and sentiment correlations,which is called the LS-stacking model,is proposed according to the factor association between multi-dimensional *** experimental results prove that the mixed feature and LS-stacking model can predict well on the 16 emotion categories of the self-built image *** study improves the fine-grained recognition ability of image emotion by computers,which helps to increase the intelligence and automation degree of visual retrieval and post-production system.

关键词： fine-grained emotion prediction movie and television scene images stacking model linear regression

来源：评论

学校读者我要写书评

暂无评论

Source Separation of Piano Concertos Using Hybrid LSTM-Transformer Model

Source Separation of Piano Concertos Using Hybrid LSTM-Trans...

引用

2024 International Conference on Culture-Oriented Science and technology, CoST 2024

作者： Liu, JingYu He, Wei Zhou, Jingjing Jiang, Wei State Key Laboratory of Media Convergence and Communication Beijing China Key Laboratory of Acoustic Visual Technology and Intelligent Control System Ministry of Culture and Tourism Beijing China School of Information and Communication Engineering Communication University of China Beijing China

ISBN: (纸本)9798350380347

Music source separation, the process of extracting independent audio streams from a complex mix, has traditionally focused on isolating vocals, drums, bass, and other primary sources. This study tackles the more intricate task of separating the piano component from a piano concerto-a challenge compounded by the diverse range of instruments and the dynamic shifts in volume and timbre. Unlike traditional music separation tasks, the piano's distinct characteristics and its interaction with the orchestra demand a more nuanced *** address the scarcity of multi-track recordings for piano concertos, this research pioneers an artificial data synthesis strategy to create a robust training dataset. We introduce a novel hybrid deep learning model that integrates Long Short-Term Memory (LSTM) networks with Transformer architecture, capitalizing on their complementary strengths to distinguish piano melodies from the rich tapestry of orchestral sounds. Our experiments demonstrate that this hybrid approach significantly outperforms conventional methods, with an improvement of 3.18 dB in signal-to-distortion ratio. These results not only validate the efficacy of proposed method but also pave the way for innovative applications in classical music source separation. © 2024 IEEE.

关键词： Source separation

来源：评论

学校读者我要写书评

暂无评论

A Singing Formant Extraction Method Based on Peak Fitting

A Singing Formant Extraction Method Based on Peak Fitting

引用

2024 International Conference on Culture-Oriented Science and technology, CoST 2024

作者： Liu, Jingyu Wu, Jiawei Jiang, Wei State Key Laboratory of Media Convergence and Communication Communication University of China 100024 China Key Laboratory of Acoustic Visual Technology and Intelligent Control System Ministry of Culture and Tourism Communication University of China 100024 China Communication University of China School of Information and Communication Engineering 100024 China

ISBN: (纸本)9798350380347

Analyzing the timbre characteristics of different singing genres is an important issue in the field of singing acoustics. Formants are an important characteristic parameter for timbre perception, but classical acoustic formant extraction methods lack accuracy and applicability in the analysis of singing genres. In order to solve the problem of insufficient accuracy and applicability of singing formants in singing genre analysis, this paper proposes a singing formant extraction method based on peak fitting. The paper summarizes the theory and methods of extracting vocal formants and based on a large amount of experimental data analysis, proposes a vocal formant feature extraction method based on peak fitting;The effectiveness of the singing formant extraction method based on peak fitting was verified through comparative experiments with existing formant algorithms. This article provides new research ideas and methods for extracting acoustic features in singing. © 2024 IEEE.

关键词： audio feature extraction formant peak fitting singing acoustic singing genre

来源：评论

学校读者我要写书评

暂无评论

Association between Timbre Perception Features and Fine-Grained Emotions in Film and Television Music

Association between Timbre Perception Features and Fine-Grai...

引用

2024 International Conference on Culture-Oriented Science and technology, CoST 2024

作者： Ren, Xiaomeng Su, Zhibin Jiang, Wei Liu, Jingyu State Key Laboratory of Media Convergence and Communication Communication University of China 100024 China Ministry of Culture and Tourism Communication University of China Key Laboratory of Acoustic Visual Technology and Intelligent Control System 100024 China School of Information and Communication Engineering Communication University of China 100024 China

ISBN: (纸本)9798350380347

As an integral component of modern cultural works, the role of film and television music in artistic expression and emotional communication has become increasingly significant. Timbre, as a crucial attribute of music, has primarily been studied in terms of the influence of its fundamental objective parameters on coarse-grained emotions. However, there is a lack of in-depth exploration of the association between timbre perception features and fine-grained emotion. Consequently, this paper approaches from the perspective of timbre perception features and fine-grained emotions in film and television music. By conducting perceptual evaluation and subjective emotional annotation experiments, it analyzes the correlation between timbre perception features and fine-grained emotions, confirming that timbre perception features significantly influence musical emotions. Furthermore, a multivariate linear regression model was employed to construct a fine-grained emotion prediction model, which has demonstrated good predictive ability for most musical emotions. This research provides a new perspective on understanding the relationship between timbre and emotion in film and television music, offering a scientific basis for its creation and performance. © 2024 IEEE.

关键词： fine-grained emotion multiple linear regression timbre emotion association timbre perception features

来源：评论

学校读者我要写书评

暂无评论

A semantic and emotion-based dual latent variable generation model for a dialogue system

引用

CAAI Transactions on Intelligence technology 2023年第2期8卷 319-330页

作者： Ming Yan Xingrui Lou Chien Aun Chan Yan Wang Wei Jiang State Key Laboratory of Media Convergence and Communication Communication University of ChinaBeijingChina School of Information and Communications Engineering Communication University of ChinaBeijingChina Key Laboratory of Acoustic Visual Technology and Intelligent Control System Communication University of ChinaBeijingChina Department of Electrical and Electronic Engineering The University of MelbourneMelbourneVictoriaAustralia School of Data Science and Intelligent Media Communication University of ChinaBeijingChina

With the development of intelligent agents pursuing humanisation,artificial intelligence must consider emotion,the most basic spiritual need in human *** emotional dialogue systems usually use an external emotional dictionary to select appropriate emotional words to add to the response or concatenate emotional tags and semantic features in the decoding step to generate appropriate ***,selecting emotional words from a fixed emotional dictionary may result in loss of the diversity and consistency of the *** propose a semantic and emotion-based dual latent variable generation model(Dual-LVG)for dialogue systems,which is able to generate appropriate emotional responses without an emotional *** from previous work,the conditional variational autoencoder(CVAE)adopts the standard transformer ***,Dual-LVG regularises the CVAE latent space by introducing a dual latent space of semantics and *** content diversity and emotional accuracy of the generated responses are improved by learning emotion and semantic features ***,the average attention mechanism is adopted to better extract semantic features at the sequence level,and the semi-supervised attention mechanism is used in the decoding step to strengthen the fusion of emotional features of the *** results show that Dual-LVG can successfully achieve the effect of generating different content by controlling emotional factors.

关键词： conditional variational autoencoder dual latent space emotional responses latent variable generation

来源：评论

学校读者我要写书评

暂无评论

An Acoustic Dataset Construction Method of Chinese Traditional Musical Instrument Guqin

An Acoustic Dataset Construction Method of Chinese Tradition...

引用

2024 International Conference on Culture-Oriented Science and technology, CoST 2024

作者： Liu, Jingyu Yang, Chen Li, Zijin Jiang, Wei State Key Laboratory of Media Convergence and Communication Communication University of China 100024 China Ministry of Culture and Tourism Communication University of China Key Laboratory of Acoustic Visual Technology and Intelligent Control System 100024 China Communication University of China School of Information and Communication Engineering 100024 China Central Conservatory of Music Ai Music and Music Information Technology Department 100032 China

ISBN: (纸本)9798350380347

With the acceleration of modernization and the rapid development of popular music, the splendid culture of Guqin, a type of Chinese traditional musical instrument which has lasted for nearly 1,000 years, is facing a serious inheritance crisis. Therefore, how to make full use of modern technical means to digitally store and manage the valuable traditional music cultural resources is the current *** response to the aforementioned challenges, this study focuses on the Guqin, an emblematic Chinese traditional musical instrument. By employing advanced audio feature extraction techniques, this research systematically analyzes the diverse playing techniques of Guqin, thereby establishing a comprehensive acoustic dataset. This endeavor not only facilitates the digital preservation and management of the Guqin's musical heritage but also pioneers a standardized framework for the creation of acoustic datasets. This framework serves as a benchmark for the development of similar datasets, thereby promoting the conservation and standardization of resources associated with Chinese traditional musical instruments. This will provide a scientific basis for the inheritance, standardized management and construction of more acoustic datasets of ethnic musical instruments in the future. © 2024 IEEE.

关键词： Musical instruments

来源：评论

学校读者我要写书评

暂无评论

Piano Concerto Source Separation Based on Speech Separation and Channel Attention

Piano Concerto Source Separation Based on Speech Separation ...

引用

2024 International Conference on Culture-Oriented Science and technology, CoST 2024

作者： Liu, Jingyu Zhou, JingJing He, Xinyan He, Wei Communication University of China State Key Laboratory of Media Convergence and Communication 100024 China Communication University of China Key Laboratory of Acoustic Visual Technology and Intelligent Control System Ministry of Culture and Tourism 100024 China Communication University of China School of Information and Communication Engineering 100024 China Communication University of China School of Faculty of International Media 100024 China

ISBN: (纸本)9798350380347

The Mossformer model excels in speech separation but has not been effectively applied to music source separation. Music sources have complex characteristics and higher sampling rates, making separation tasks more challenging. We addressed a rarely explored task of separating piano concerto recordings into individual piano and orchestral tracks. This process involves intricate coordination between the piano and orchestra, creating highly complex audio signals in both time and frequency domains. Our main contributions include: (1) adapting the speech separation model for the novel task of piano concerto source separation, constructing and processing a specialized dataset.(2) introducing channel attention in the separation module to dynamically adjust feature focus based on instrument characteristics, enhancing key features. Experiments on the Piano Concerto Dataset (PCD) showed improved separation performance, with a 0.22dB average Signal-to-Distortion Ratio (SDR) increase over the baseline model. © 2024 IEEE.

关键词： Source separation

来源：评论

学校读者我要写书评

暂无评论

A Classification and Recognition Model of Distributed Resources Based on Feature Extraction

A Classification and Recognition Model of Distributed Resour...

引用

2023 IEEE International Conference on Energy Technologies for Future Grids, ETFG 2023

作者： Zhu, Yu Yang, Le Yang, Xizai Lu, Jiaxin Wang, Yuqing Wang, Fei State Grid Shaanxi Electric Power Co. Ltd. Information and Communication Company Information and Communication Technology R & D Center Xi'an710048 China State Key Laboratory of Power System Operation and Control Tsinghua University Beijing100084 China

ISBN: (纸本)9781665471640

The extensive integration of distributed resources such as distributed photovoltaic system and electric vehicles enhances the uncertainty of the load side greatly. Corresponding demand response strategies can be developed for different distributed resource types and characteristics of households' load, which is of great significance for new energy consumption and stable operation of the power system. However, only the net load power of households can be obtained through smart meters instead of the distributed resource types of households. Based on this, a distributed resource classification and identification model based on feature extraction is proposed in this paper, and effectively solves the problems of low accuracy and strong data dependence of existing recognition methods. Firstly, a generalized weather class generation algorithm based on clustering and voting is established, which is used to determine the weather class of each day to provide a basis for the following feature extraction. Then, based on the typical net load profiles under different generalized weather classes, a two-stage feature extraction and a classification identification model based on integrated learning algorithms are established to identify whether a user contains DPVS and EV, respectively. Finally, based on the category labels of each user for each stage, the final classification of the category to which the user belongs is carried out. Simulation experiments show that the recognition model built using the typical features extracted in the paper has good recognition accuracy. © 2023 IEEE.

关键词： classification and recognition Distributed photovoltaics system electric vehicles ensemble learning feature extraction

来源：评论

学校读者我要写书评

暂无评论

The Regional Distributed PV Ultra-short-term Power Forecasting Based on Static-dynamic Spatiotemporal Correlation Modeling

The Regional Distributed PV Ultra-short-term Power Forecasti...

引用

2023 IEEE International Conference on Energy Technologies for Future Grids, ETFG 2023

作者： Zhu, Yu Yang, Le Yang, Xizai Xu, Fei Su, Zinuo Zhen, Zhao Information and Communication Technology R & D Center State Grid Shaanxi Electric Power Co. Ltd. Information and Communication Company Xi'an710048 China State Key Laboratory of Power System Operation and Control Tsinghua University Beijing100084 China

ISBN: (纸本)9781665471640

Accurate prediction of distributed photovoltaic power can help the arrange the scheduling plan reasonably to Maintain the safe and stable operation of the distribution network. However, distributed photovoltaic power generation exists lack of meteorological data and real-time monitoring data, the distribution range is large and dispersed, the installed capacity is relatively small and so on, the centralized photovoltaic power prediction method is not applicable to distributed photovoltaic power prediction, so it is necessary to carry out a regional power prediction study for distributed photovoltaic. Distributed PV is located in close proximity, in a similar external environment, and the power output between stations has strong spatial and temporal correlation, which can be captured to improve the accuracy of power forecasting. To this end, a regional distributed PV power prediction model based on static-dynamic spatiotemporal correlation modeling is proposed. The k-means is firstly used to cluster the regional distributed photovoltaic sites into several sub-regions, according to the geographical location and output characteristics that reflect the long-term static correlation between power stations. Then, graph convolutional neural network are applied to capture the dynamic correlation between subregions according to the power. In this way, the static and dynamic correlation are effectively considered in power prediction modeling, and the prediction accuracy has been improved. The superiority of the proposed method is verified by a real world dataset in China. © 2023 IEEE.

关键词： distributed photovoltaic graph convolution Regional power prediction

来源：评论

学校读者我要写书评

暂无评论

Scalable parallel ultrafast optical random bit generation based on a single chaotic microcomb

引用

Light(Science & Applications) 2024年第4期13卷 637-644页

作者： Pu Li Qizhi Li Wenye Tang Weiqiang Wang Wenfu Zhang Brent ELittle Sai Tek Chu KAlan Shore Yuwen Qin Yuncai Wang Institute of Advanced Photonics Technology School of Information EngineeringGuangdong University of TechnologyGuangzhou 51006China. Key Laboratory of Photonic Technology for Integrated Sensing and Communication Ministry of Education of ChinaGuangdong University of TechnologyGuangzhou 51006China. Guangdong Provincial Key Laboratory of Information Photonics Technology Guangdong University of TechnologyGuangzhou 51006China. Key Laboratory of Advanced Transducers andIntelligent Control System Ministry of EducationTaiyuan University of TechnologyTaiyuan 030024China. State Key Laboratory of Transient Optics and Photonics Xi’an Institute of Optics and Precision MechanicsChinese Academy of SciencesXi’an 710119China. Department of Physics and Materials Science City University of Hong KongHong KongChina. School of Electronic Engineering Bangor UniversityBangorWales LL571UTUK

Random bit generators are critical for information security,cryptography,stochastic modeling,and *** and scalability are key challenges faced by current physical random bit ***,we propose a massively parallel scheme for ultrafast random bit generation towards rates of order 100 terabit per second based on a single micro-ring resonator.A modulation-instability-driven chaotic comb in a micro-ring resonator enables the simultaneous generation of hundreds of independent and unbiased random bit streams.A proof-of-concept experiment demonstrates that using our method,random bit streams beyond 2 terabit per second can be successfully generated with only 7 comb *** bit rate can be easily enhanced by further increasing the number of comb lines *** approach provides a chip-scale solution to random bit generation for secure communication and high-performance computation,and offers superhigh speed and large scalability.

关键词： resonator random chaotic

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：