检索结果-内蒙古大学图书馆

A knowledge-based deep learning method for ECG signal delineation

FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE 2020年 109卷 56-66页

作者： Wang, Jilong Li, Renfa Li, Rui Fu, Bin Hunan Univ Coll Comp Sci & Elect Engn Changsha 410082 Hunan Peoples R China Hunan Univ Key Lab Embedded & Network Comp Hunan Prov Changsha 410082 Hunan Peoples R China

The delineation of electrocardiograms (ECG) is a crucial step designed to extract signal characteristics and assist cardiologists in diagnosing certain diseases. It refers to the delineation of both the onset and offset of the P wave, the QRS complex, and the T wave. However, to date previous studies have neither investigated the role of medical knowledge nor investigated the interference of signal noise. In this paper, we propose a knowledge-based deep learning method for ECG signal delineation, which adds domain knowledge and individual feature knowledge to improve delineation performance. The proposed method is novel in three distinct ways. First, the method aligns the encoded knowledge with ECG signals with reference to the R wave peak position. Second, the method incorporates domain knowledge to delineate ECG signals under the encoder-decoder framework. Third, the method introduces individual feature knowledge to adjust the delineation results adaptively. An evaluation conducted on the QT database demonstrates that the proposed method can obtain, on average, high performance with sensitivity of 99.62% and positive predictivity of 99.81%. The evaluation also shows that the method can obtain a sensitivity and a positive predictivity above 90% in most noisy cases. (C) 2020 Elsevier B.V. All rights reserved.

关键词： Deep learning Domain knowledge ECG delineation encoder-decoder model Knowledge encoding

来源：评论

学校读者我要写书评

暂无评论

ConvLSTM Based Temperature Forecast Modification model for North China

引用

Journal of Tropical Meteorology 2022年第4期28卷 405-412页

作者： GENG Huan-tong HU Zhong-yan WANG Tian-lei School of Computer Science Nanjing University of Information Science and TechnologyNanjing 210044 China Jiangsu Open University Nanjing 210036 China

The correction of model forecast is an important step in evaluating weather forecast *** recent years,post-processing models based on deep learning have become *** this paper,a deep learning model named EDConvLSTM based on encoder-decoder structure and ConvLSTM is developed,which appears to be able to effectively correct numerical weather *** with traditional post-processing methods and convolutional neural networks,ED-ConvLSTM has strong collaborative extraction ability to effectively extract the temporal and spatial features of numerical weather forecasts and fit the complex nonlinear relationship between forecast field and observation *** this paper,the post-processing method of ED-ConvLSTM for 2 m temperature prediction is tested using The International Grand Global Ensemble dataset and ERA5-Land data from the European Centre for Medium-Range Weather Forecasts(ECMWF).Root mean square error and temperature prediction accuracy are used as evaluation indexes to compare ED-ConvLSTM with the method of model output statistics,convolutional neural network postprocessing methods,and the original prediction by the *** results show that the correction effect of EDConvLSTM is better than that of the other two postprocessing methods in terms of the two indexes,especially in the long forecast time.

关键词： temperature forecast post-processing numerical weather prediction encoder-decoder model ConvLSTM

来源：评论

学校读者我要写书评

暂无评论

ReFIGG: retinal fundus image generation using GAN

引用

INTERNATIONAL JOURNAL OF COMPUTATIONAL SCIENCE AND ENGINEERING 2023年第3期26卷 316-323页

作者： Nair, Sharika Sasidharan Meharban, M. S. Rajagiri Sch Engn & Technol Dept Comp Sci & Engn Ernakulam Kerala India

The effective training of deep architectures mainly depends on a large number of well-explained data. This is a problem in the medical field where it is hard and costly to gain such images. The tiny blood-vessels of the retina are the only part of the human structure that can be directly and non-intrusively foreseen within the living. Hence, it can be easily obtained and examined by automatic tools. Fundus imaging;a basic check-up process in ophthalmology provides essential data to make it easier for doctors to detect various eye-related diseases at prior phases. Fundus image generation is a challenging process to carry out by constructing composite models of the eye structure. In this paper, we overcome the issue of unavailability of medical fundus datasets by synthesising them artificially through an encoder-decoder generator model to the existing MCML method of generative adversarial networks (GAN) for easier, quicker, and early analysis.

关键词： fundus image generative adversarial networks GAN encoder-decoder model image synthesis deep learning

来源：评论

学校读者我要写书评

暂无评论

An attention-based hybrid deep learning approach for bengali video captioning

引用

JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES 2023年第1期35卷 257-269页

作者： Zaoad, Md. Shahir Mannan, M. M. Rushadul Mandol, Angshu Bikash Rahman, Mostafizur Islam, Md Adnanul Rahman, Md. Mahbubur Mil Inst Sci & Technol Dept Comp Sci & Engn Dhaka 1216 Bangladesh

Video captioning is an automated process of captioning a video by understanding the content within it. Although numerous studies have been performed on video captioning in English, the field of video cap-tioning in Bengali remains nearly unexplored. Therefore, this research aims at generating Bengali cap-tions that plausibly describe the gist of a specific video as well as identifying the best performing model for Bengali video captioning. To accomplish this, several sequence-to-sequence models - LSTM, BiLSTM, and GRU are implemented that takes the video frame features as input, extracted through differ-ent CNN models - VGG-19, Inceptionv3, and ResNet50v2, and provides a corresponding textual descrip-tion as output. Moreover, the Attention mechanism is incorporated with these models as a first-ever attempt in Bengali video captioning. In this study, a novel Bengali video captioning dataset is constructed from Microsoft Research Video Description Corpus (MSVD) dataset (an English video captioning dataset) through utilizing a deep learning-based translator and manual post-editing efforts. Finally, the model's performance is evaluated in terms of popular performance evaluation metrics -BLEU, METEOR, and ROUGE. The proposed attention-based hybrid model outperforms the existing models in terms of these evaluation metrics, establishing a new benchmark for Bengali video captioning. (c) 2022 The Author(s). Published by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY-NC-ND license (http://***/licenses/by-nc-nd/4.0/).

关键词： Bengali video captioning Convolutional neural network encoder-decoder model Recurrent neural network Attention-mechanism

来源：评论

学校读者我要写书评

暂无评论

mmArrhythmia: Contactless Arrhythmia Detection via mmWave Sensing

引用

PROCEEDINGS OF THE ACM ON INTERACTIVE MOBILE WEARABLE AND UBIQUITOUS TECHNOLOGIES-IMWUT 2024年第1期8卷 1-25页

作者： Zhao, Langcheng Lyu, Rui Lin, Qi Zhou, Anfu Zhang, Huanhuan Ma, Huadong Wang, Jingjia Shao, Chunli Tang, Yida Beijing Univ Posts & Telecommun 10 Xitucheng Rd Beijing Peoples R China Peking Univ Third Hosp 49 Huayuan North Rd Beijing Peoples R China

Arrhythmia is a common problem of irregular heartbeats, which may lead to serious complications such as stroke and even mortality. Due to the paroxysmal nature of arrhythmia, its long-term monitoring and early detection in daily household scenarios, instead of depending on ECG examination only available during clinical visits, are of critical importance. While ambulatory ECG Holter and wearables like smartwatches have been used, they are still inconvenient and interfere with users' daily activities. In this paper, we bridge the gap by proposing mmArrhythmia, which employs low-cost mmWave radar to passively sense cardiac motions and detect arrhythmia, in an unobtrusive contact-less way. Different from previous mmWave cardiac sensing works focusing on healthy people, mmArrhythmia needs to distinguish the minute and transient abnormal cardiac activities of arrhythmia patients. To overcome the challenge, we custom-design an encoder-decoder model that can perform arrhythmia feature encoding, sampling and fusion over raw IQ sensing data directly, so as to discriminate normal heartbeat and arrhythmia. Furthermore, we enhance the robustness of mmArrhythmia by designing multichannel ensemble learning to solve the model bias problem caused by unbalanced arrhythmia data distribution. Empirical evaluation over 79,910 heartbeats demonstrates mmArrhythmia's ability of robust arrhythmia detection, with 97.32% accuracy, 98.63% specificity, and 92.30% sensitivity.

关键词： Contactless arrhythmia detection millimeter wave sensing encoder-decoder model ensemble learning

来源：评论

学校读者我要写书评

暂无评论

Using Auto-encoder BiLSTM Neural Network for Czech Grapheme-to-Phoneme Conversion 22nd

Using Auto-Encoder BiLSTM Neural Network for Czech Grapheme-...

引用

22nd Annual International Conference on Text, Speech, and Dialogue (TSD)

作者： Juzova, Marketa Vit, Jakub Univ West Bohemia Fac Appl Sci Dept Cybernet & New Technol Informat Soc Plzen Czech Republic

ISBN: (纸本)9783030279479;9783030279462

The crucial part of almost all current TTS systems is a grapheme-to-phoneme (G2P) conversion, i.e. the transcription of any input grapheme sequence into the correct sequence of phonemes in the given language. Unfortunately, the preparation of transcription rules and pronunciation dictionaries is not an easy process for new languages in TTS systems. For that reason, in the presented paper, we focus on the creation of an automatic G2P model, based on neural networks (NN). But, contrary to the majority of related works in G2P field, using only separate words as an input, we consider a whole phrase the input of our proposed NN model. That approach should, in our opinion, lead to more precise phonetic transcription output because the pronunciation of a word can depend on the surrounding words. The results of the trained G2P model are presented on the Czech language where the cross-word-boundary phenomena occur quite often, and they are compared to the rule-based approach.

关键词： Grapheme-to-phoneme Sequence-to-sequence Neural networks encoder-decoder model Czech phonetic transcription

来源：评论

学校读者我要写书评

暂无评论

Transformer-Based Unified Neural Network for Quality Estimation and Transformer-Based Re-decoding model for Machine Translation 16th

Transformer-Based Unified Neural Network for Quality Estimat...

引用

16th China Conference on Machine Translation (CCMT)

作者： Chen, Cong Zong, Qinqin Luo, Qi Qiu, Bailian Li, Maoxi Jiangxi Normal Univ Nanchang Jiangxi Peoples R China

ISBN: (纸本)9789813361621;9789813361614

In this paper, we describe our submitted system for CCMT 2020 sentence-level quality estimation subtasks and machine translation subtasks. We propose two models: (i) a Transformer-based unified neural network for the quality estimation of machine translation, which consists of a Transformer bottleneck layer and a bidirectional long short-term memory network that are jointly optimized and fine-tuned for quality estimation, and (ii) a Transformer-based re-decoding model for machine translation, which takes the generated machine translation outputs as the approximate contextual environment of the target language and synchronously re-decodes each token in the machine translation outputs. Experimental results on the development set show that the proposed approaches outperform the baseline systems.

关键词： Machine translation Quality estimation of machine translation Re-decoding encoder-decoder model

来源：评论

学校读者我要写书评

暂无评论

Low-Resource Speech Synthesis with Speaker-Aware Embedding 13

Low-Resource Speech Synthesis with Speaker-Aware Embedding

引用

13th International Symposium on Chinese Spoken Language Processing (ISCSLP)

作者： Yang, Li-Jen Yeh, I-Ping Chien, Jen-Tzung Natl Yang Ming Chiao Tung Univ Inst Elect & Comp Engn Hsinchu Taiwan Natl Yang Ming Chiao Tung Univ Grad Degree Program Cybersecur Hsinchu Taiwan

ISBN: (纸本)9798350397963

Speech synthesis has been successfully exploited for mapping from text sequence to speech waveform where high-resource languages have been well studied and learned from a large amount of text-speech paired data in public-domain corpora. However, developing speech synthesis under low-resource languages is challenging for speech communication in local regions since the collection of training data is expensive. In particular, the speaker-aware speech generation under low-resource settings is crucial in real world. Such a problem is increasingly difficult in case of very limited speaker-specific data. This paper presents a speaker-aware speech synthesis under low-resource settings based on an encoder-decoder framework by using transformer. Knowledge transfer is performed by incorporating a speaker-aware embedding through first learning a pretrained transformer from multi-speaker data of a low-populated spoken language and then fine-tuning the transformer to a target speaker with very limited speaker-specific embeddings. Experiments on low-resource Taiwanese speech synthesis are evaluated to show the merit of speaker-aware transformer in terms of Mel cepstral distortion and mean opinion score.

关键词： low-resource speech synthesis speaker-aware embedding encoder-decoder model transformer

来源：评论

学校读者我要写书评

暂无评论

Language modeling in speech recognition for grammatical error detection based on neural machine translation

引用

ACOUSTICAL SCIENCE AND TECHNOLOGY 2020年第5期41卷 788-791页

作者： Fu, Jiang Chiba, Yuya Nose, Takashi Ito, Akinori Tohoku Univ Grad Sch Engn Aoba Ku Aramaki Aza Aoba 6-6-05 Sendai Miyagi 9808579 Japan

来源：评论

学校读者我要写书评

暂无评论

Handwritten Mathematical Expression Recognition with Bidirectionally Trained Transformer 16th

Handwritten Mathematical Expression Recognition with Bidirec...

引用

16th IAPR International Conference on Document Analysis and Recognition (ICDAR)

作者： Zhao, Wenqi Gao, Liangcai Yan, Zuoyu Peng, Shuai Du, Lin Zhang, Ziyin Peking Univ Wangxuan Inst Comp Technol Beijing Peoples R China Huawei AI Applicat Res Ctr Beijing Peoples R China

ISBN: (纸本)9783030863319

encoder-decoder models have made great progress on handwritten mathematical expression recognition recently. However, it is still a challenge for existing methods to assign attention to image features accurately. Moreover, those encoder-decoder models usually adopt RNN-based models in their decoder part, which makes them inefficient in processing long LATEX sequences. In this paper, a transformer-based decoder is employed to replace RNN-based ones, which makes the whole model architecture very concise. Furthermore, a novel training strategy is introduced to fully exploit the potential of the transformer in bidirectional language modeling. Compared to several methods that do not use data augmentation, experiments demonstrate that our model improves the ExpRate of current state-of-the-art methods on CROHME 2014 by 2.23%. Similarly, on CROHME 2016 and CROHME 2019, we improve the ExpRate by 1.92% and 2.28% respectively.

关键词： Handwritten mathematical expression recognition Transformer Bidirection encoder-decoder model

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：