检索结果-内蒙古大学图书馆

Intelligent Technologies (CONIT), International Conference on

作者： Anand Vinod Neethu Mohan Sachin Kumar S Soman K P Center for Computational Engineering and Networking (CEN) Amrita Vishwa Vidyapeetham Coimbatore India

The COVID-19 pandemic engulfed the entire world. RT-PCR assessment nowadays is a metric golden standard for contemplating COVID. However, this method takes time and violates social distancing and thus necessary steps have to be taken for its prediction in the future due to its communicable property. Cough is an important bio-marker and it contains micro patterns and audio fingerprints that can be used for classification purposes. Thus in this paper, the main focus is given to the cough data for the classification of COVID-19 diseases. The necessary features such as ZCR, MFCC, chroma STFT, roll-off, spectral centroid, and spectral bandwidth are extracted from the cough data. Traditional machine learning models as well as various deep learning models are implemented in order to find out the model with maximum accuracy for COVID-19 identification.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Fast Iterative Filtering Method for Efficient Denoising of Phonocardiogram Signals

A Fast Iterative Filtering Method for Efficient Denoising of...

引用

Intelligent Technologies (CONIT), International Conference on

作者： N Sai Satwik Reddy V Poorna Muni Saisdhar Reddy Venkata Siva Manoj A Neethu Mohan Sachin Kumar S Soman K P Center for Computational Engineering and Networking (CEN) Amrita Vishwa Vidyapeetham Coimbatore India

The need for computer-aided diagnosis (CAD) in the automatic monitoring of cardiac diseases has significantly increased in the current medical arena. This paper presents a novel denoising approach based on Fast Iterative Filtering (FIF) for CAD in the automatic monitoring of cardiac diseases. FIF efficiently eliminates unwanted noise from phonocardiogram (PCG) signals, preserving information for precise diagnosis. The comparative analysis highlights the exceptional performance of the FIF method in enhancing cardiac signal denoising tasks in automated heart monitoring systems. The proposed framework integrates a robust phonocardiography system with suitable hardware and a computer, utilizing the FIF technique to denoise PCG signals. The system provides a visual representation of the denoised waveform to the user and sends cardiac information to physicians for further analysis.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Sentiment Analysis at Document level of Telugu Data from Multi-Domains

Sentiment Analysis at Document level of Telugu Data from Mul...

引用

Intelligent Technologies (CONIT), International Conference on

作者： Katipally Vighneshwar Reddy Sachin Kumar S KP Soman Center for Computational Engineering and Networking (CEN) Amrita Vishwa Vidyapeetham Coimbatore India

Finding common threads of optimism or negative in text is the goal of sentiment analysis. In the business world, it is used to learn about a product's reception, a customer's identity, and their expectations of a firm by monitoring the tone of online conversations on forums like Reddit and Twitter. Sentiment analysis in various languages has arisen as a separate topic of NLP as businesses seek for client feedback in previously untapped markets. The stakes could not be greater, since Telugu is a Dravidian language spoken by almost 82 million people. Little effort, such as annotated data and software, is put towards supporting Telugu, hence the language is often overlooked. To better understand the sentiment of the data, we conducted our research using the "Sentirama" dataset and, as you'll see in the paper, we used a variety of Machine Learning models (including SVM-linear, SVM-quadratic, SVM-polynomial, Random Forest, Naive Bayes, and KNN) and Featured concepts (including word2vec+ (CBOW or skip-gram), TF-IDF, and Fastext). To find the best Deep-learning model for Telugu sentiment analysis, we also tried out LSTM, Bidirectional-LSTM, and 1D-CNN.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-Focus Image Fusion using 2D Compact Variational Mode Decomposition

Multi-Focus Image Fusion using 2D Compact Variational Mode D...

引用

Intelligent Technologies (CONIT), International Conference on

作者： Swarnamuhi K A Swetha Tirumalai Neethu Mohan Sachin Kumar S Soman K.P Center for Computational Engineering and Networking (CEN) Amrita Vishwa Vidyapeetham Coimbatore India

Fusion is the technique to integrate noteworthy information from different devices such as sensors using mathematical techniques. The major motive of image fusion is to generate an image that has maximized information and clarity while maintaining integrity. Image fusion aims to reduce redundant data and to construct image more eligible and comprehensible for perception. Fusion is now an indispensable part of technical and industrial applications. This paper investigate the significance of 2D Compact Variational Mode Decomposition (2D-CVMD) algorithm for multi focus fusion of image. The methodology proposed emphasises on multi-focus images. Through extensive experiments it is validated that the proposed methodology generates final fused image with maximum information and better explanation of the view than the original images. Further the accuracy of estimation is validated using standard quality metrics on different multi-focus fusion images.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Learning (With) Distributed Optimization

arXiv

引用

arXiv 2023年

作者： Aadhithya, Aadharsh A. Abinesh, S. Akshaya, J. Jayanth, M. Radhakrishnan, Vishnu Sowmya, V. Soman, K.P. Center for Computational Engineering and Networking School of Computing Amrita Vishwa Vidyapeetham Coimbatore India

This paper provides an overview of the historical progression of distributedoptimization techniques, tracing their development from early duality-basedmethods pioneered by Dantzig, Wolfe, and Benders in the 1960s to the emergenceof the Augmented Lagrangian Alternating Direction Inexact Newton (ALADIN)algorithm. The initial focus on Lagrangian relaxation for convex problems anddecomposition strategies led to the refinement of methods like the AlternatingDirection Method of Multipliers (ADMM). The resurgence of interest indistributed optimization in the late 2000s, particularly in machine learningand imaging, demonstrated ADMM\'s practical efficacy and its unifying *** overview also highlights the emergence of the proximal center method andits applications in diverse domains. Furthermore, the paper underscores thedistinctive features of ALADIN, which offers convergence guarantees fornon-convex scenarios without introducing auxiliary variables, differentiatingit from traditional augmentation techniques. In essence, this work encapsulatesthe historical trajectory of distributed optimization and underscores thepromising prospects of ALADIN in addressing non-convex optimization challenges. Copyright © 2023, The Authors. All rights reserved.

关键词： Constrained optimization

来源：评论

学校读者我要写书评

暂无评论

Age-Based Automatic Voice Conversion Using Blood Relation for Voice Impaired

引用

Computers, Materials & Continua 2022年第2期70卷 4027-4051页

作者： Palli Padmini C.Paramasivam G.Jyothish Lal Sadeen Alharbi Kaustav Bhowmick Department of Electronics&Communication Engineering Amrita School of EngineeringBengaluruAmrita Vishwa VidyapeethamIndia Center for Computational Engineering and Networking(CEN) Amrita School of EngineeringCoimbatoreAmrita Vishwa VidyapeethamIndia Department of Software Engineering College of Computer and Information SciencesKing Saud UniversityRiyadhSaudi Arabia Department of Electronics and Communication Engineering PES UniversityBengaluruIndia

The present work presents a statistical method to translate human voices across age groups,based on commonalities in voices of blood *** age-translated voices have been naturalized extracting the blood relation features e.g.,pitch,duration,energy,using Mel Frequency Cepstrum Coefficients(MFCC),for social compatibility of the *** system has been demonstrated using standard English and an Indian *** voice samples for resynthesis were derived from 12 families,with member ages ranging from 8–80 *** voice-age translation,performed using the Pitch synchronous overlap and add(PSOLA)approach,by modulation of extracted voice features,was validated by perception *** translated and resynthesized voices were correlated using Linde,Buzo,Gray(LBG),and Kekre’s Fast Codebook generation(KFCG)*** translated voice targets,a strong(θ>∼93%andθ>∼96%)correlation was found with blood relatives,whereas,a weak(θ<∼78%andθ<∼80%)correlation range was found between different families and different gender from same *** study further subcategorized the sampling and synthesis of the voices into similar or dissimilar gender groups,using a support vector machine(SVM)choosing between available voice ***,∼96%,∼93%,and∼94%accuracies were obtained in the identification of the gender of the voice sample,the age group samples,and the correlation between the original and converted voice samples,*** results obtained were close to the natural voice sample features and are envisaged to facilitate a near-natural voice for speech-impaired easily.

关键词： Blood relations KFCG LBG MFCC vector quantization correlation speech samples same-gender dissimilar gender voice conversion PSOLA SVM

来源：评论

学校读者我要写书评

暂无评论

A Real-Time Oral Cavity Gesture Based Words Synthesizer Using Sensors

引用

Computers, Materials & Continua 2022年第6期71卷 4523-4554页

The present system experimentally demonstrates a synthesis of syllables and words from tongue manoeuvers in multiple languages,captured by four oral sensors *** an experimental demonstration of the system used in the oral cavity,a prototype tooth model was *** on the principle developed in a previous publication by the author(s),the proposed system has been implemented using the oral cavity(tongue,teeth,and lips)features alone,without the glottis and the *** positions of the sensors in the proposed system were optimized based on articulatory(oral cavity)gestures estimated by simulating the mechanism of human *** system has been tested for all English alphabets and several words with sensor-based input along with an experimental demonstration of the developed algorithm,with limit switches,potentiometer,and flex sensors emulating the tongue in an artificial oral *** system produces the sounds of vowels,consonants,and words in English,along with the pronunciation of meanings of their translations in four major Indian languages,all from oral cavity *** experimental setup also caters to gender mapping of *** sound produced from the hardware has been validated by a perceptual test to verify the gender and word of the speech sample by listeners,with∼98%and∼95%accuracy,*** a model may be useful to interpret speech for those who are speech-disabled because of accidents,neuron disorder,spinal cord injury,or larynx disorder.

关键词： English vowels and consonants oral cavity proposed system sensors speech-disabled speech production vocal tract model

来源：评论

学校读者我要写书评

暂无评论

Grapheme to Phoneme Mapping for Tamil Language 5th

Grapheme to Phoneme Mapping for Tamil Language

引用

5th International Conference on Advances in Computing and Data Sciences, ICACDS 2021

作者： Anusha, M. Geerthana Govind, D. Menon, Vijay Krishna Center for Computational Engineering and Networking Amrita School of Engineering Amrita Vishwa Vidyapeetham Coimbatore India

ISBN: (纸本)9783030882433

In this paper, a Deep learning-based approach is analyzed to improve the accuracy of the Language Modelling of a system that is a part of ASR. The main intent is to build an End to End Neural Network Language Model. Generally, ASR works in three different modes. The modes are Acoustic Modeling, Lexicons, and Language Modeling. In ASR, G2P models are vital in the case of analyzing the Out Of Vocabulary words. Mostly these G2P models were fabricated using the traditional approaches such as Bi-gram, n-gram language models. The traditional approaches are deceptive in aiding the meanings, long term dependencies of the words and are also impotent in rendering the phonemes of new words. Thus Seq2Seq G2P models exceed the state of the art of the traditional approach. As we all know, the ASR (Speech to Text) system is widely used in AI applications such as voice dialing, system control, navigation, etc. The method which is applied here is the Statistical based one. A method other than the Rule-based approach is trialed out. A suitable neural network of LSTM and BI-LSTM is proposed which enhances the accuracy of the system. © 2021, Springer Nature Switzerland AG.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Comparative Study of Bert Models and Roberta in Transformer based Question Answering

Comparative Study of Bert Models and Roberta in Transformer ...

引用

Intelligent Technologies (CONIT), International Conference on

作者： N Akhila Sanjanasri J. P Soman K. P Center for Computational Engineering and Networking (CEN) Amrita School of Engineering Coimbatore Amrita Vishwa Vidyapeetham Coimbatore India

Using deep learning technique, transformer-based self-supervised pre-trained models have revolutionised the idea of transfer learning in natural language processing (NLP). The self-attention mechanism makes transformers more prevalent in transfer learning across broad range of NLP tasks. In this study, seven prominent models are compared., including the Bertbase- uncased, Distilbert-base-cased, Distilbert-base-uncased, and Roberta models, in terms of their effectiveness (using 3 epochs). Bert base cased, Bert-medium-squad2-distilled, and Electra-basesquad2 on the Stanford Question Answering Dataset were utilised models with two epochs (SQuAD). The analysis shows that in three epochs, Roberta models provide the most accuracy, while Distilbert-base cased models provide the maximum accuracy when compared to Bert-base uncased, Distilbert-base uncased, and Distilbert base cased models. While Electra-base-squad2 performs better than Bert base cased and Bert-medium-squad2- distilled in cases with two epoches. Although all Bert and Roberta models take a long time to run, increasing accuracy requires more time to train the dataset.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Performance Improvement of Deep Residual Skip Convolution Neural Network for Atrial Fibrillation Classification 8th

Performance Improvement of Deep Residual Skip Convolution Ne...

引用

8th International Conference on Frontiers of Intelligent Computing: Theory and Applications, FICTA 2020

作者： Sanjana K Sowmya, V. Gopalakrishnan, E.A. Soman, K.P. Amrita School of Engineering Center for Computational Engineering and Networking Amrita Vishwa Vidyapeetham CoimbatoreTamilNadu India

ISBN: (纸本)9789811557873

Atrial fibrillation is a life-threatening cardiac disease which requires a long and tedious process of detection. So, the detection of atrial fibrillation has gained great importance. One of the most reliable ways to detect cardiac disease is through analysis of ECG signal. In this paper, we show that the performance of a deep residual skip convolution neural network-based approach for automatic detection of atrial fibrillation can be improved by hyperparameter tuning. For the present work, atrial fibrillation dataset from the 2017 PhysioNet/CinC Challenge is used. The proposed method obtained an overall accuracy of 96.08% and weighted average F1 score of 0.96, a recall of 0.96 and a precision of 0.96. The main advantage of the present work is the improved accuracy achieved using a lighter model which is trained for a lesser number of epochs. © 2021, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Convolution

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：