检索结果-内蒙古大学图书馆

5th IEEE Global Conference for Advancement in Technology, GCAT 2024

作者： Santhosh, Abhijith Rohith, K.P. Ananthakrishnan, S. Rajesh, Aravind Bhaskar, Jasmine Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering Amritapuri India

ISBN: (纸本)9798350376685

Visual Question Answering is a multi-modal task to find an answer to any question asked about a given image. In this paper, we propose a model of dual CNNs that provide the correct answer to a spoken question based on a given image in audio or text form. Mirroring real-world scenarios like helping the visually impaired, the questions and answers are considered in spoken mode. Visual questions selectively address different areas of an image, including underlying context and background information. In this paper, we introduce a dual-CNN model - a conjunction of two convolutional neural networks concerning Visual Question Answering. This first CNN model is used to select the significant features from the image that contribute to the Answer, and the second CNN to extract relevant textual features from the question. The proposed framework has been tested with ***. In terms of overall accuracy and execution time, the experimental findings demonstrated very good results. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Enhanced Aerial Vehicle Detection Using a Customized YOLOv8 Model for Real-World Applications 9

Enhanced Aerial Vehicle Detection Using a Customized YOLOv8 ...

引用

9th International Conference on Communication and Electronics Systems, ICCES 2024

作者： Thalapilly, Maanav Agasti, Ketone Nambiar, Niranjan S. Suresh, Ananthu Anjali, T. Amrita School of Computing Amrita Vishwa Vidyapeetham Department of Computer Science and Engineering Amritapuri India

ISBN: (纸本)9798350377972

Aerial vehicle detection is a critical function in surveillance, disaster response, and many defense systems. In this paper, a customized model, which extends the YOLOv8 model to detect aerial vehicles with exact precision and efficiency with four classes-categorized as helicopters, airplanes, drones, and UAVs-is developed. The dataset comprises 5,050 labeled images of various sources, augmented for robustness, and formatted in the YOLO-compatible structure for smooth training. The proposed approach integrates optimized depth and width multipliers for reducing complexity of computation at a preserved high detection accuracy. Its efficiency reflects in an mAP of 98.7%, which is superior to YOLOv8 (97.1%) and YOLOv5 (95.8%). Comparison in the two regimes indicates that the work offers superior precision and computational efficiency, hence suitable for real-world scenarios where resources are constrained. Results confirm the proposed approach is efficient in dealing with the cases of small objects, overlapping entities, and complex backgrounds, and its potential could be for further development of aerial surveillance technologies. © 2024 IEEE.

关键词： Aircraft detection

来源：评论

学校读者我要写书评

暂无评论

News Topic Categorization on Telugu Dataset Using Deep Learning Models 15

News Topic Categorization on Telugu Dataset Using Deep Learn...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Deepika, M. Sree Subramanyam, N. Sri Lekhya Pranav, P. Shiva Reddy, Pulaputhuru Subramanyam Venugopalan, Manju Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering Bengaluru India

ISBN: (纸本)9798350370249

The digital age has ushered in an unprecedented era of data proliferation across various domains, with the news sector being no exception, accumulating vast amounts of information, often measured in petabytes. Manual classification of such voluminous data is impractical, particularly considering the complexity of languages like Telugu, belonging to the Dravidian family, renowned for its intricate morphological characteristics. Natural Language Processing (NLP) offers a promising avenue for tackling textual data, leveraging diverse word embedding techniques to enrich model representations. Notably, prior research in this domain has predominantly relied on count-based algorithms for word embeddings. However, in this investigation, we explore alternative approaches, specifically custom-built Deep Learning models integrated with word embeddings, to optimize the classification of newspaper articles. Through a series of methodical experiments, we endeavor to identify the most effective model architecture for this task, aiming to enhance the accuracy and efficiency of article categorization in Telugu. The proposed model has experimented with word embeddings such as FastText and GloVE with deep learning models like LSTM, Bi-LSTM, CNN. Among the models experimented, FastText in combination with CNN have reported the highest F1-Score of 0.93. © 2024 IEEE.

关键词： Categorization Deep Learning News data Word Embedding

来源：评论

学校读者我要写书评

暂无评论

Automatic Assessment of Quadratic Equation Solutions Using MathBERT and RoBERTa Embeddings 15

Automatic Assessment of Quadratic Equation Solutions Using M...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Rao, Subhiksha Suresh Mishra, Sahil Akhilesh, S. Balakrishnan, Roshni M. Pati, Peeta Basa Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science & Engineering Bengaluru India

ISBN: (纸本)9798350370249

This project aims to create a machine learning classification system that classifies solutions to quadratic equations into three different groups: 'Incorrect,' 'Partially Correct,' and 'Correct.' MathBERT and RoBERTa embeddings are utilized in this system. A multi-class classification technique is used by the system, with an emphasis on feature engineering, data preparation, and model selection. To improve model performance, methods like Principal Component Analysis (PCA), LIME (Local Interpretable Model-Agnostic Explanations), and ensemble meta models with Boosting, Voting, and Stacking are used. The efficacy of many classification models in correctly classifying quadratic solutions is examined in this work. To evaluate the performance of the model, evaluation measures including F1-score, accuracy, precision, and recall are used. The results show that MathBERT embeddings outperform RoBERTa embeddings. Ensemble models, especially those that include SVM and MLP classifiers, achieve the best performance, with the stacking approach yielding an accuracy of approximately 65%. This model accurately classifies solutions into the correct categories, demonstrating the potential for practical applications in mathematics education. © 2024 IEEE.

关键词： Automated Evaluation Classification Educational Technology Machine Learning MathBERT Quadratic Equations RoBERTa Solution Accuracy

来源：评论

学校读者我要写书评

暂无评论

Advancing Gender, Age and Ethnicity with YOLOv5 and Transfer Learning 6th

Advancing Gender, Age and Ethnicity with YOLOv5 and Transfer...

引用

6th International Conference on Information Systems and Management science, ISMS 2023

作者： Bhattacharya, Varun Tripathy, Balakrushna School of Computer Science Engineering and Information Systems Vellore Institute of Technology Vellore632014 India

ISBN: (纸本)9783031664090

The detection of gender, age, and ethnicity using machine learning techniques holds significant promise for applications in marketing, healthcare, and law enforcement. Previous methods relied on conventional machine learning and manual feature engineering, which often struggled to capture the complex relationships between facial features and these demographic factors. To overcome these limitations, our novel approach leverages the powerful YOLOv5 framework, integrating transfer learning and deep neural networks. This cutting-edge method excels at accurately classifying gender, age, and ethnicity from facial images, thanks to its deep neural network architecture with advanced object detection capabilities. Our work represents a substantial advancement in gender, age, and ethnicity prediction, showcasing the potential of advanced machine learning techniques. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Securing Neural Network-Based Personalized Medicine for Advanced Liver Cancer Detection in Healthcare

Securing Neural Network-Based Personalized Medicine for Adva...

引用

2024 IEEE International Conference on Information Technology, Electronics and Intelligent Communication Systems, ICITEICS 2024

作者： Kasturi, Salvadi Singh, Jagendra Agarwal, Mohit Kumar, Ashwini Mishra, Komal Sharma, Yogita Cvr College of Engineering Department of Computer Science Engineering Vastunagar India School of Computer Science Engineering & Technology Bennett University Greater Noida India Compucom Institute of Technology & Management Department of Computer Science Engineering Jaipur India Chandigarh University Department of University Institute of Computing Chandigarh India School of Management and Commerce Manav Rachna University Faridabad India

ISBN: (纸本)9798350382693

This research is aimed at discussing the critical meeting point of healthcare and artificial intelligence, including the need to protect the required treatment solution. Investigation into the utility of various deep learning techniques such as VGG 16, VGG 19, Convolutional Neural Networks (CNN), and Recurrent Neural Network (RNN) for the prediction of liver cancer stages and their respective treatments has been completed. The generation of 2340 of medical images from multiple sources, including the internet and records of hospitals, as a dataset was performed in a secure cloud. The security measures have allowed only authorized personnel to access the data and ensured the privacy and secrecy of data. The created dataset was then separated into two parts: training, which consisted of training the model on 70% of images, and testing, which implied that 30% of images were used to test the generated model. For each of the run models, performance evaluation measurements such as accuracy, precision, recall, and F1 score were calculated. The results have shown the highest efficiency of VGG 19 model, which resulted in 98.44% of accuracy. For the generated model, a fictitious ROC curve and specificity rate were also provided to showcase the discrimination abilities of each of the submitted models. The results show that the utility of neural network-based personalized medicine is beneficial for the further progression of the early prediction stage of liver cancer and its treatment. The result showed that the proposed models has successfully contributed to the field of healthcare informatics and further opportunities for artificial intelligence application to increase patient outcomes in the form of personalized medicine that may be performed. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Protein Sequence Classification Using Bidirectional Encoder Representations from Transformers (BERT) Approach

引用

SN computer science 2023年第5期4卷 481页

作者： Balamurugan, R. Mohite, Saurabh Raja, S.P. School of Computer Science and Engineering Vellore Institute of Technology Tamilnadu Vellore 632014 India

Proteins play a vital role by booming out a number of activities within an organism to withstand its life. The field of Natural Language Processing has successfully adapted deep learning to get a better insight into the semantic nature of languages. In this paper, we propose semantic approaches based on deep learning to work with protein amino acid sequences and compare the performances of these approaches with traditional classifiers to predict their respective families. The Bidirectional Encoder Representations from Transformers (BERT) approach was tested over 103 protein families from UniProt consortium database. The results show the average prediction accuracy to 99.02%, testing accuracy to 97.70%, validation accuracy to 97.69%, Normalized Mutual Information (NMI) score on overall data to 98.45, on test data to 96.99, on validation data 96.93 with high weighted average F1 scores of 99.02 on overall data, 97.72 on test data and 97.70 on validation data, and high macro average F1 scores of 99.00 on overall data, 98.00 on test data and 98.00 on validation data. From the results, it is justified that our proposed approach is outperforming well when compared to the existing approaches. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Document embedding Logistic Regression Natural Language Processing Protein family

来源：评论

学校读者我要写书评

暂无评论

Detection of Real and Manipulated Videos using the Transfer Learning Approach of Deep-Learning Models 5

Detection of Real and Manipulated Videos using the Transfer ...

引用

5th IEEE Global Conference for Advancement in Technology, GCAT 2024

作者： Reddy, Bandi Rupendra Rup, Daka Chandra Rohith, Mathi Belwal, Meena Amrita Vishwa Vidyapeetham Amrita School of Computing Department of Computer Science and Engineering Bengaluru India

ISBN: (纸本)9798350376685

Deepfake (manipulated images and videos) identification is critical in multimedia forensics, as it can spread misinformation, threaten privacy, and have legal implications. To address this problem, researchers and practitioners are developing deepfake detection techniques that use deep-learning models and computer vision algorithms to analyse the authenticity of images and videos. This study proposes a novel approach providing robust and scalable deepfake detection by using transfer learning in feature extraction and deep-learning models to train features extracted from the video dataset. The transfer-learning has been experimented with the ResNet, XceptionNet, and VGG16 models for feature extractions. Further, the deep-learning models Long short-term memory (LSTM) and Bidirectional long short-term memory (Bi-LSTM) have been experimented with to train the extracted features. To rigorously test the proposed approach, the data set under the experiment combined the three benchmark datasets - The Deepfake Detection Challenge (DFDC), Celeb-DF, and Face-Forensics++ datasets. The prediction rate achieved was 96 % and above on average by the proposed approach for the datasets under consideration for the experiments. The proposed approach performs better than the state-of-the-art techniques, gaining a higher accuracy of 96% and above. © 2024 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Sea Surface Temperature Prediction using Sample Convolution and Interaction Network 15

Sea Surface Temperature Prediction using Sample Convolution ...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Gayathri, P. Yalavarthi, Rahul Varma, Jayanth Santosh Gopakumar, G. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Amritapuri India

ISBN: (纸本)9798350370249

Predicting oceanic phenomena plays a very important role in ensuring maritime safety, managing coastal areas and preparing them for any disasters like cyclone or tsunami. There are several factors that contribute to these severe weather events among which Sea Surface Temperature(SST) is a crucial one. Warmer oceans can often lead to stronger storms or even frequent heavy downpours, highlighting the importance of accurately forecasting SST. This study explores the performance of a deep learning model called Sample Convolution and Interaction Network (SCINet) in forecasting SST. This model follows a decompose-downsample-convolve-interact architecture that helps in learning complex trends in the data. On comparing SCINet’s performance with well established time-series models like Long-Short Term Memory (LSTM) network, it was able to achieve almost 4% - 20% lower Mean Absolute Error (MAE) than LSTM on various testing points. This highlights SCINet’s potential effectiveness in time-series forecasting, contributing to the ongoing development and revolutionary advancements in deep learning models for time series prediction. ©2024 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

YoloV8-Powered Surveillance: Safeguarding Crops from Wildlife and Fire Hazards 15

YoloV8-Powered Surveillance: Safeguarding Crops from Wildlif...

引用

15th International Conference on Computing Communication and Networking Technologies, ICCCNT 2024

作者： Nishanth, K. Pranav, Shri S. Jayasuriyaa, E. Lalithamani, N. Prasand, Ajey M. Department of Computer Science and Engineering Amrita School of Computing Amrita Vishwa Vidyapeetham Coimbatore India

ISBN: (纸本)9798350370249

Agriculture is the backbone of India's economy. However, the farmers are facing crop damage due to wild animals encounter and fire. The study has mainly intended to provide a solution for the prevention of wildlife encounters into agricultural field and fire using Object Detection. This study mainly focuses on detection of fire and four distinct animals: Elephant, Wild Boar, Deer and Cow. The dataset contains 7000 images with annotated five categories of Object. We trained five YOLOv8 models (nano - YOLOv8n, small - YOLOv8s, medium - YOLOv8m, large - YOLOv8l, extra-large – YOLOv8x) on our dataset. Raspberry Pi equipped with a Camera captures video of the crop field and applies Yolov8 (Object Detection Algorithm) to detect fire and wildlife encounters and plays some loud noises to deter wild animals. The Farmer is notified with a photograph of Farm and class name of the animal detection made through a Telegram Bot. The trained medium YOLOv8m model has attained a mAP of 92.2%, proving its efficacy in identifying fire and wild animals. By integrating this device across farm field, real-time automated monitoring and alerts can be provided to farmers, minimizing crop losses. © 2024 IEEE.

关键词： Fire hazards

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：