检索结果-内蒙古大学图书馆

Speech Enhancement Based on Discrete Wavelet Packet Transform and Itakura-Saito nonnegative matrix factorisation

ARCHIVES OF ACOUSTICS 2020年第4期45卷 565-572页

作者： Liu, Houguang Wang, Wenbo Xue, Lin Yang, Jianhua Wang, Zhihua Hua, Chunli China Univ Min & Technol Sch Mechatron Engn Xuzhou 221116 Jiangsu Peoples R China

nonnegative matrix factorization (NMF) is one of the most popular machine learning tools for speech enhancement (SE). However, there are two problems reducing the performance of the traditional NMF-based SE algorithms. One is related to the overlap-and-add operation used in the short time Fourier transform (STFT) based signal reconstruction, and the other is the Euclidean distance used commonly as an objective function;these methods can cause distortion in the SE process. In order to get over these shortcomings, we propose a novel SE joint framework which combines the discrete wavelet packet transform (DWPT) and the Itakura-Saito nonnegative matrix factorisation (ISNMF). In this approach, the speech signal was first split into a series of subband signals using the DWPT. Then, the ISNMF was used to enhance the speech for each subband signal. Finally, the inverse DWPT (IDWT) was utilised to reconstruct these enhanced speech subband signals. The experimental results show that the proposed joint framework effectively enhances the performance of speech enhancement and performs better in the unseen noise case compared to the traditional NMF methods.

关键词： speech enhancement discrete wavelet packet transform nonnegative matrix factorisation Itakura-Saito divergence

来源：评论

学校读者我要写书评

暂无评论

SUPERVISED GROUP nonnegative matrix factorisation WITH SIMILARITY CONSTRAINTS AND APPLICATIONS TO SPEAKER IDENTIFICATION

SUPERVISED GROUP NONNEGATIVE MATRIX FACTORISATION WITH SIMIL...

引用

IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Serizel, Romain Bisot, Victor Essid, Slim Richard, Gael Univ Lorraine LORIA UMR 7503 F-54506 Vandoeuvre Les Nancy France INRIA F-54600 Villers Les Nancy France CNRS LORIA UMR 7503 F-54506 Vandoeuvre Les Nancy France Univ Paris Saclay Telecom ParisTech LTCI CNRS F-75013 Paris France

ISBN: (纸本)9781509041176

This paper presents supervised feature learning approaches for speaker identification that rely on nonnegative matrix factorisation. Recent studies have shown that group nonnegative matrix factorisation and task-driven supervised dictionary learning can help performing effective feature learning for audio classification problems. This paper proposes to integrate a recent method that relies on group nonnegative matrix factorisation into a task-driven supervised framework for speaker identification. The goal is to capture both the speaker variability and the session variability while exploiting the discriminative learning aspect of the task-driven approach. Results on a subset of the ESTER corpus prove that the proposed approach can be competitive with I-vectors.

关键词： nonnegative matrix factorisation feature learning dictionary learning online learning speaker identification

来源：评论

学校读者我要写书评

暂无评论

Supervised group nonnegative matrix factorisation with similarity constraints and applications to speaker identification

Supervised group nonnegative matrix factorisation with simil...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Romain Serizel Victor Bisot Slim Essid Gael Richard Universite de Lorraine LORIA UMR 7503 Vandauvre-les-Nancy F-54506 France LTCI CNRS Telecom ParisTech Universite Paris - Saclay F-75013 France

ISBN: (纸本)9781509041183

关键词： nonnegative matrix factorisation feature learning dictionary learning online learning speaker identification

来源：评论

学校读者我要写书评

暂无评论

GROUP nonnegative matrix factorisation WITH SPEAKER AND SESSION VARIABILITY COMPENSATION FOR SPEAKER IDENTIFICATION 41

GROUP NONNEGATIVE MATRIX FACTORISATION WITH SPEAKER AND SESS...

引用

41st IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

作者： Serizel, Romain Essid, Slim Richard, Gael Univ Paris Saclay CNRS LTCI Telecom ParisTech F-75013 Paris France

ISBN: (纸本)9781479999880

This paper presents a feature learning approach for speaker identification that is based on nonnegative matrix factorisation. Recent studies have shown that with such models, the dictionary atoms can represent well the speaker identity. The approaches proposed so far focused only on speaker variability and not on session variability. However, this later point is a crucial aspect in the success of the I-vector approach that is now the state-of-the-art in speaker identification. This paper proposes a method that relies on group nonnegative matrix factorisation and that is inspired by the I-vector training procedure. By doing so the proposed approach intends to capture both the speaker variability and the session variability. Results on a small corpus prove that the proposed approach can be competitive with I-vectors.

关键词： nonnegative matrix factorisation spectrogram factorisation feature learning speaker variability speaker identification

来源：评论

学校读者我要写书评

暂无评论

GROUP nonnegative matrix factorisation WITH SPEAKER AND SESSION VARIABILITY COMPENSATION FOR SPEAKER IDENTIFICATION

GROUP NONNEGATIVE MATRIX FACTORISATION WITH SPEAKER AND SESS...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Romain Serizel Slim Essid Gael Richard LTCI CNRS Télécom ParisTech Université Paris - Saclay 75013 Paris France

ISBN: (纸本)9781479999897

关键词： nonnegative matrix factorisation spectrogram factorisation feature learning speaker variability speaker identification speaker recognition Loudspeakers matrix decomposition compensated

来源：评论

学校读者我要写书评

暂无评论

ULTRASOUND-COUPLED SEMI-SUPERVISED nonnegative matrix factorisation FOR SPEECH ENHANCEMENT

ULTRASOUND-COUPLED SEMI-SUPERVISED NONNEGATIVE MATRIX FACTOR...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

作者： Barker, Tom Virtanen, Tuomas Delhomme, Olivier Tampere Univ Technol Dept Signal Proc FIN-33101 Tampere Finland

ISBN: (纸本)9781479928934

We present an extension to an existing speech enhancement technique, whereby the incorporation of easily obtained Doppler-based ultrasound data, obtained from frequency shifts caused by a talker's mouth movements, is shown to improve speech enhancement results. Noisy speech mixtures were enhanced using semi-supervised nonnegative matrix factorisation (NMF). Ultrasound data recorded alongside the speech is transformed into the spectral domain and used additionally to audio in the mixture to be separated. Speech components are learned from a training set, whilst noise components are estimated from the mixture signal. We show that the ultrasound data can improve source-to-distortion ratios for the enhanced speech, relative to both the non-ultrasound NMF case and an established Wiener filter-based speech enhancement method.

关键词： nonnegative matrix factorisation Ultrasound Acoustic Doppler Sensor Source Separation

来源：评论

学校读者我要写书评

暂无评论

Efficient calibration of high-dimensional dynamic OD matrix in metropolitan networks: combining dimensionality reduction and bagging-based metamodel

引用

TRANSPORTMETRICA A-TRANSPORT SCIENCE 2025年

作者： Peng, Chang Xu, Chengcheng Wang, Changshuai Tong, Hao Ren, Weilin Southeast Univ Sch Transportat Si Pai Lou 2 Nanjing 210096 Peoples R China

This paper presents an efficient and easily implemented dynamic origin-destination (OD) calibration framework that integrates dimensionality reduction and metamodel techniques to address high dimensionality and computational inefficiency. A spatial-temporal nonnegative matrix factorization (ST-NMF) algorithm was developed to capture basic patterns in OD flows, reducing the solution space. Two novel and easy-implemented bagging-based metamodels were developed for calibration acceleration by replacing a time-consuming simulation testbed with multiple low-complexity simulators in gradient estimation. Each simulator is built on randomly sampled OD flows to guarantee computational efficiency. The framework was evaluated in a real-world metropolitan area, where a dynamic OD matrix with nearly 3.5 million variables was calibrated in minimal iterations by reducing the number of decision variables to 90. Compared to a benchmark using only metamodels, it reduced simulation errors by 30-40%. Against the benchmarks using only dimensionality reduction, it saved over 40% of computational time.

关键词： Dynamic OD calibration high-dimensionality nonnegative matrix factorisation metamodel bagging method

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Survey on Multi-View Clustering

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2023年第12期35卷 12350-12368页

作者： Fang, Uno Li, Man Li, Jianxin Gao, Longxiang Jia, Tao Zhang, Yanchun Deakin Univ Fac Sci Engn & Built Environm Sch IT Geelong Vic 3220 Australia Qilu Univ Technol Shandong Acad Sci Shandong Comp Sci Ctr Natl Supercomp Ctr Jinan Jinan 400700 Shandong Peoples R China Southwest Univ Sch Informat & Technol Chongqing 400700 Peoples R China Guangzhou Univ Cyberspace Inst Adv Technol New Cyber Res Dept Peng Cheng Lab China Guangzhou 510006 Guangdong Peoples R China

The development of information gathering and extraction technology has led to the popularity of multi-view data, which enables samples to be seen from numerous perspectives. Multi-view clustering (MVC), which groups data samples by leveraging complementary and consensual information from several views, is gaining popularity. Despite the rapid evolution of MVC approaches, there has yet to be a study that provides a full MVC roadmap for both stimulating technical improvements and orienting research newbies to MVC. In this article, we review recent MVC techniques with the purpose of exhibiting the concepts of popular methodologies and their advancements. This survey not only serves as a unique MVC comprehensive knowledge for researchers but also has the potential to spark new ideas in MVC research. We summarise a large variety of current MVC approaches based on two technical mechanisms: heuristic-based multi-view clustering (HMVC) and neural network-based multi-view clustering (NNMVC). We end with four technological approaches within the category of HMVC: nonnegative matrix factorisation, graph learning, latent representation learning, and tensor learning. Deep representation learning and deep graph learning are two technical methods that we demonstrate in NNMVC. We also show 15 publicly available multi-view datasets and examine how representative MVC approaches perform on them. In addition, this study identifies the potential research directions that may require further investigation in order to enhance the further development of MVC.

关键词： Multi-view clustering graph learning nonnegative matrix factorisation deep representation learning contrastive learning

来源：评论

学校读者我要写书评

暂无评论

Graph regularised sparse NMF factorisation for imagery de-noising

引用

IET COMPUTER VISION 2018年第4期12卷 466-475页

作者： Fang, Yixian Zhang, Huaxiang Ren, Yuwei Shandong Normal Univ Sch Informat Sci & Engn Jinan 250014 Shandong Peoples R China Qilu Univ Technol Shandong Acad Sci Sch Sci Jinan 250353 Shandong Peoples R China Shandong Normal Univ Inst Data Sci & Technol Jinan 250014 Shandong Peoples R China

When utilising non-negative matrix factorisation (NMF) to decompose a data matrix into the product of two low-rank matrices with non-negative entries, the noisy components of data may be introduced into the matrix. Many approaches have been proposed to address the problem. Different from them, the authors consider the group sparsity and the geometric structure of data by introducing L-2,L-1-norm and local structure preserving regularisation in the formulated objective function. A graph regularised sparse NMF de-noising approach is proposed to learn discriminative representations for the original data. Since the non-differentiability of L-2,L-1-norm increases the computational cost, they propose an effective iterative multiplicative update algorithm to solve the objective function by using the Frobenius-norm of transpose coefficient matrix. Experimental results on facial image datasets demonstrate the superiority of the proposed approach over several state-of-the-art approaches.

关键词： image denoising matrix decomposition graph theory sparse matrices image representation iterative methods face recognition learning (artificial intelligence) graph regularised sparse NMF factorisation imagery denoising nonnegative matrix factorisation data matrix decomposition two low-rank matrices product group sparsity 2 1-norm local structure preserving regularisation graph regularised sparse NMF denoising approach discriminative representation iterative multiplicative update algorithm transpose coefficient matrix Frobenius-norm facial image datasets

来源：评论

学校读者我要写书评

暂无评论

Homotopy optimisation based NMF for audio source separation

引用

IET SIGNAL PROCESSING 2018年第9期12卷 1099-1106页

作者： Koundinya, Sriharsha Karmakar, Abhijit CEERI CSIR Integrated Syst Grp Pilani Rajasthan India Acad Sci & Innovat Res Madras Tamil Nadu India

In this study, the authors propose a novel framework for audio source separation based on a cascaded non-negative matrix factorisation (NMF) using homotopy optimisation with perturbation and ensemble (HOPE) and denoising autoencoder. NMF using traditional optimisation has a problem of finding a global solution, and hence could not achieve complete separation of the sources from the mixture. This problem has been addressed using homotopy optimisation in this study. Subsequently, using denoising autoencoder the residual sounds that are usually observed in the separated sources are removed. The enhanced audio signals are filtered using Wiener techniques to obtain the separated signals. The homotopy-based NMF is applied for separating singing voice and drums from song samples using a single-channel mixture. The separated signals are compared with other NMF algorithms by using Blind Source Separation (BSS) Eval objective quality measures. The NMF with HOPE and denoising autoencoder is shown to provide an improvement of up to 6dB in comparison with other NMF algorithms.

关键词： audio signal processing blind source separation source separation matrix decomposition denoising autoencoder traditional optimisation complete separation separated sources enhanced audio signals separated signals homotopy-based NMF separating singing voice drums NMF algorithms homotopy optimisation-based NMF audio source separation nonnegative matrix factorisation noise figure 6 0 dB

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：