检索结果-内蒙古大学图书馆

48th IEEE International Conference on Acoustics, speech and Signal processing, ICASSP 2023

作者： Shah, Zehra Qi, Shi-Ang Wang, Fei Farrokh, Mahtab Tasnim, Mashrura Stroulia, Eleni Greiner, Russell Plitsis, Manos Katsamanis, Athanasios University of Alberta Department of Computing Science Edmonton Canada Athena Research Center Institute for Language and Speech Processing Greece

ISBN: (纸本)9781728163277

We explore ways to use speech data to screen for indications of Alzheimer's dementia (AD). In particular, we describe our approach to the ICASSP 2023 Signal processing Grand Challenge, which involves extrapolating from models learned from English speech samples, to Greek speech samples, to determine which subjects have AD. By using acoustic and linguistic features, inspired by clinical research on AD, our top-performing classification model achieves 69% accuracy in distinguishing AD patients from healthy controls, and our regression model attains an RMSE of 4.8 for inferring cognitive testing scores. These outcomes underscore the potential of our explainable model for detecting cognitive decline in AD patients via speech, and its applicability in clinical settings. © 2023 IEEE.

关键词： Alzheimer's dementia Machine Learning Signal processing speech

来源：评论

学校读者我要写书评

暂无评论

Adapted Multimodal Bert with Layer-Wise Fusion for Sentiment Analysis 48

Adapted Multimodal Bert with Layer-Wise Fusion for Sentiment...

引用

48th IEEE International Conference on Acoustics, speech and Signal processing, ICASSP 2023

作者： Chlapanis, Odysseas S. Paraskevopoulos, Georgios Potamianos, Alexandros National Technical University of Athens Athens Greece Institute for Language and Speech Processing Athena Research Center Athens Greece

ISBN: (纸本)9781728163277

Multimodal learning pipelines have benefited from the success of pretrained language models. However, this comes at the cost of increased model parameters. In this work, we propose Adapted Multimodal BERT (AMB), a BERT-based architecture for multimodal tasks that uses a combination of adapter modules and intermediate fusion layers. The adapter adjusts the pretrained language model for the task at hand, while the fusion layers perform task-specific, layer-wise fusion of audio-visual information with textual BERT representations. During the adaptation process the pre-trained language model parameters remain frozen, allowing for fast, parameter-efficient training. In our ablations we see that this approach leads to efficient models, that can outperform their fine-tuned counterparts and are robust to input noise. Our experiments on sentiment analysis with CMU-MOSEI show that AMB outperforms the current state-of-the-art across metrics, with 3.4% relative reduction in the resulting error and 2.1% relative improvement in 7-class classification accuracy. © 2023 IEEE.

关键词： adapters BERT fusion multimodal

来源：评论

学校读者我要写书评

暂无评论

The EASIER Mobile Application and Avatar End-User Evaluation Methodology 11

The EASIER Mobile Application and Avatar End-User Evaluation...

引用

11th Workshop on the Representation and processing of Sign languages: Evaluation of Sign language Resources, sign-lang@LREC-COLING 2024

作者： Picron, Frankie Van Landuyt, Davy Omardeen, Rehana Efthimiou, Eleni Wolfe, Rosalee Fotinea, Stavroula-Evita Goulas, Theodore Tismer, Christian Kopf, Maria Hanke, Thomas European Union of the Deaf Belgium Institute for Language and Speech Processing ATHENA Research Center Greece Nuromedia GmbH Germany Institute of German Sign Language and Communication of the Deaf University of Hamburg Germany

ISBN: (纸本)9782493814302

Here we report on the methodological approach adopted for the end-user evaluation studies carried out during the lifecycle of the EASIER project, focusing on the project's mobile app and avatar technologies. Evaluation was led by deaf consortium partners and performed in two cycles, involving both deaf signers and hearing sign language (SL) experts groups from five SLs to provide user feedback, which served as a reference to base the next development steps of the respective EASIER components. With this goal in mind, priorities were (i) to exploit information gathered via focus group discussions after (ii) presenting evaluators with the technological components and related questionnaires fully accessible to signers to maximize feedback and underline the importance of user involvement in the development of the technology. © 2024 ELRA language Resources Association: CC BY-NC 4.0.

关键词： Audition

来源：评论

学校读者我要写书评

暂无评论

Preparing Ultrasound Imaging Data for Artificial Intelligence Tasks: Anonymisation, Cropping, and Tagging

Preparing Ultrasound Imaging Data for Artificial Intelligenc...

引用

18th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2025

作者： Pechlivanis, Dimitrios Didaskalou, Stylianos Kaldoudi, Eleni Drosatos, George School of Medicine Democritus University of Thrace Alexandroupoli68100 Greece Institute for Language and Speech Processing Athena Research Center Xanthi67100 Greece

Ultrasound imaging is a widely used diagnostic method in various clinical contexts, requiring efficient and accurate data preparation workflows for artificial intelligence (AI) tasks. Preparing ultrasound data presents challenges such as ensuring data privacy, extracting diagnostically relevant regions, and associating contextual metadata. This paper introduces a standalone application designed to streamline the preparation of ultrasound DICOM files for AI applications across different medical use cases. The application facilitates three key pro- cesses: (1) anonymisation, ensuring compliance with privacy standards by removing sensitive metadata;(2) cropping, isolating relevant regions in images or video frames to enhance the utility for AI analysis;and (3) tagging, enriching files with additional metadata such as anatomical position and imaging purpose. Built with an intuitive interface and robust backend, the application optimises DICOM file processing for efficient inte- gration into AI workflows. The effectiveness of the tool is evaluated using a dataset of Deep Vein Thrombosis (DVT) ultrasound images, demonstrating significant improvements in data preparation efficiency. This work establishes a generalizable framework for ultrasound imaging data preparation while offering specific insights into DVT-focused AI workflows. Future work will focus on further automation and expanding support to ad- ditional imaging modalities as well as evaluating the tool in a clinical setting. © 2025 by SCITEPRESS – Science and Technology Publications, Lda.

关键词： Ultrasonic imaging

来源：评论

学校读者我要写书评

暂无评论

Data management in Ethereum DApps: A cost and performance analysis

Data management in Ethereum DApps: A cost and performance an...

引用

作者： Kostamis, Periklis Sendros, Andreas Efraimidis, Pavlos S. Department of Electrical and Computer Engineering Democritus University of Thrace Xanthi67100 Greece Institute for Language and Speech Processing Athena Research Center Xanthi67100 Greece

The cost of interacting with a blockchain and the time required to search and retrieve information from it are important factors in the design of Decentralized Applications (DApps). The corresponding approaches for data management in Ethereum vary from pure on-chain solutions to hybrid architectures. In this work, we thoroughly examine a wide range of data management methods, including unconventional ones, and evaluate them in terms of storage cost and retrieval latency while considering the impact of recent hard forks. More precisely, we elaborate on the use of Smart Contract (SC) storage as a data store for DApps and in view of the prohibitive cost inherent to this approach, we present and compare methods for low-cost data handling in Ethereum, namely the event-logs, the transaction payload, and the almost surprising exploitation of unused function arguments. In addition, we study hybrid schemes by using IPFS and Swarm as storage platforms and Ethereum as a timestamping proof mechanism. Such schemes are especially effective when large chunks of data have to be managed. The upload and retrieval performance of these platforms and the cost associated with storing (on-chain) the identifiers of the uploaded data are of utmost importance and therefore are investigated in detail. © 2023

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Prototype based Masked Audio Model for Self-Supervised Learning of Sound Event Detection

Prototype based Masked Audio Model for Self-Supervised Learn...

引用

2025 IEEE International Conference on Acoustics, speech, and Signal processing, ICASSP 2025

作者： Cai, Pengfei Song, Yan Jiang, Nan Gu, Qing McLoughlin, Ian National Engineering Research Center of Speech and Language Information Processing University of Science and Technology of China China ICT Cluster Singapore Institute of Technology Singapore

ISBN: (纸本)9798350368741

A significant challenge in sound event detection (SED) is the effective utilization of unlabeled data, given the limited availability of labeled data due to high annotation costs. Semi-supervised algorithms rely on labeled data to learn from unlabeled data, and the performance is constrained by the quality and size of the former. In this paper, we introduce the Prototype based Masked Audio Model (PMAM) algorithm for self-supervised representation learning in SED, to better exploit unlabeled data. Specifically, semantically rich frame-level pseudo labels are constructed from a Gaussian mixture model (GMM) based prototypical distribution modeling. These pseudo labels supervise the leaning of a Transformer-based masked audio model, in which binary cross-entropy loss is employed instead of the widely used InfoNCE loss, to provide independent loss contributions from different prototypes, which is important in real scenarios in which multiple labels may apply to unsupervised data frames. A final stage of fine-tuning with just a small amount of labeled data yields a very high performing SED model. On like-for-like tests using the DESED task, our method achieves a PSDS1 score of 62.5%, surpassing current state-of-the-art models and demonstrating the superiority of the proposed technique. © 2025 IEEE.

关键词： masked audio model prototype self-supervised learning sound event detection

来源：评论

学校读者我要写书评

暂无评论

Stargan-vc Based Cross-Domain Data Augmentation for Speaker Verification 48

Stargan-vc Based Cross-Domain Data Augmentation for Speaker ...

引用

48th IEEE International Conference on Acoustics, speech and Signal processing, ICASSP 2023

作者： Hu, Hang-Rui Song, Yan Zhang, Jian-Tao Dai, Li-Rong McLoughlin, Ian Zhuo, Zhu Zhou, Yu Li, Yu-Hong Xue, Hui University of Science and Technology of China National Engineering Research Center for Speech and Language Information Processing Hefei China Alibaba Group China

ISBN: (纸本)9781728163277

Automatic speaker verification (ASV) faces domain shift caused by the mismatch of intrinsic and extrinsic factors, such as recording device and speaking style, in real-world applications, which leads to severe performance degradation. Since single-speaker multi-condition (SSMC) data is difficult to collect in practice, existing domain adaptation methods are hard to ensure the feature consistency of the same class but different domains. To this end, we propose a cross-domain data generation method to obtain a domain-invariant ASV system. Inspired by voice conversion (VC) task, a StarGAN based generative model first learns cross-domain mappings from SSMC data, and then generates missing domain data for all speakers, thus increasing the intra-class diversity of the training set. Considering the difference between ASV and VC task, we renovate the corresponding training objectives and network structure to make the adaptation task-specific. Evaluations on achieve a relative performance improvement of about 5-8% over the baseline in terms of minDCF and EER, outperforming the CNSRC winner's system of the equivalent scale. © 2023 IEEE.

关键词： Data Augmentation Domain Adaptation Speaker Verification StarGAN

来源：评论

学校读者我要写书评

暂无评论

Improved G723.1 Codec speech Quality Under Burst Packet Loss Conditions 5th

Improved G723.1 Codec Speech Quality Under Burst Packet Loss...

引用

5th International Conference on Electrical Engineering and Control Applications, ICEECA 2022

作者： Bakri, Adil Mahdjane, Karima Amrouche, Abderrahmane Krobba, Ahmed Scientific Research and Technical Center for the Development of Arabic Language CRSTDLA Algiers Algeria Speech Communication and Signal Processing Laboratory Faculty of Electronics and Computer Science USTHB Algiers Algeria

ISBN: (纸本)9789819747757

In this paper, a Packet Loss Concealment (PLC) algorithm is proposed for G723.1 CELP-type speech coders in order to improve the quality of decoded speech in VoIP under burst packet loss. The original PLC method implemented in this codec is based on the generation of excitations. The generated excitations are used as input to the LP synthesis filter based on past LP coefficients. In an easier way, it relies on the last frame, which is well received to generate a new frame that replaces the lost frame. The proposed PLC is designed as a PLC algorithm for G.723.1. The proposed method performance is compared with the PLC algorithm employed in G.723.1, through the Perceptual Evaluation of speech Quality (PESQ). The experiments have shown that the proposed PLC algorithm provides significant enhancement of the speech quality than the PLC of G.723.1, especially in the presence of burst packet loss and voice onset conditions. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Packet loss

来源：评论

学校读者我要写书评

暂无评论

The USTC System for EEG-Music Emotion Recognition Challenge

The USTC System for EEG-Music Emotion Recognition Challenge

引用

2025 IEEE International Conference on Acoustics, speech, and Signal processing, ICASSP 2025

作者： Chen, Jiaxin Wang, Yiming Liu, Yin-Long Feng, Rui Yuan, Jiahong Ling, Zhen-Hua National Engineering Research Center of Speech and Language Information Processing University of Science and Technology of China Hefei China Interdisciplinary Research Center for Linguistic Sciences University of Science and Technology of China Hefei China

ISBN: (纸本)9798350368741

This paper presents the Neural Harmony team's submission to Task 1 (Person Identification) of the ICASSP 2025 EEG-Music Emotion Recognition Challenge, which aims to identify the subject from a given EEG segment. To enhance performance, we propose a novel architecture incorporating the Multiscale ConvBlock and integrating attention mechanisms with convolutional networks. We also reprocessed the data and trained multiple models with different train-validation splits, which were ensembled during testing to further improve robustness. Our final results on the test data exceed the challenge baseline, achieving 100% accuracy in Person Identification. Additionally, unseen subjects were introduced to evaluate the model's generalization ability, and the results confirm the model's strong adaptability to new subjects. © 2025 IEEE.

关键词： attention convolutional network EEG Person Identification

来源：评论

学校读者我要写书评

暂无评论

Sample-Efficient Unsupervised Domain Adaptation of speech Recognition Systems: A Case Study for Modern Greek

引用

IEEE/ACM Transactions on Audio speech and language processing 2024年 32卷 286-299页

作者： Paraskevopoulos, Georgios Kouzelis, Theodoros Rouvalis, Georgios Katsamanis, Athanasios Katsouros, Vassilis Potamianos, Alexandros National Technical University of Athens Graduate School of Electrical and Computer Engineering Athens10682 Greece Athena Research Center Institute for Speech and Language Processing Marousi15125 Greece National Technical University of Athens Faculty of Electrical and Computer Engineering Athens10682 Greece

Modern speech recognition systems exhibit rapid performance degradation under domain shift. This issue is especially prevalent in data-scarce settings, such as low-resource languages, where the diversity of training data is limited. In this work, we propose M2DS2, a simple and sample-efficient fine-tuning strategy for large pre-trained speech models, based on mixed source and target domain self-supervision. We find that including source domain self-supervision stabilizes training and avoids mode collapse of the latent representations. For evaluation, we collect HParl, a 120-hour speech corpus for Greek, consisting of plenary sessions in the Greek Parliament. We merge HParl with two popular Greek corpora to create GREC-MD, a test-bed for multi-domain evaluation of Greek ASR systems. In our experiments, we find that, while other Unsupervised Domain Adaptation baselines fail in this resource-constrained environment, M2DS2 yields significant improvements for cross-domain adaptation, even when only a few hours of in-domain audio are available. When we relax the problem in a weakly supervised setting, we find that independent adaptation for audio using M2DS2 and language using simple LM augmentation techniques is particularly effective, yielding word error rates comparable to the fully supervised baselines. © 2014 IEEE.

关键词： speech recognition

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：