检索结果-内蒙古大学图书馆

A novel transfer deep learning method for detection and classification of plant leaf disease

Journal of Ambient Intelligence and Humanized Computing 2023年第9期14卷 12407-12424页

作者： Kaur, Prabhjot Harnal, Shilpi Gautam, Vinay Singh, Mukund Pratap Singh, Santar Pal Chitkara University Institute of Engineering and Technology Chitkara University Punjab India School of Computer Science Engineering and Technology Bennett University Greater Noida India Department of Computer Science and Engineering Rashtrakavi Ramdhari Singh Dinkar College of Engineering Begusarai India

The major cause of plant mortality and devastation, particularly among trees is plant diseases. This problem, however, may be handled and treated effectively through early detection. Crop/plant diseases must be identified and prevented as soon as possible to improve yield. Plant disease classification and identification by plant protection professionals through eye observation is a time-consuming and error-prone process. Previously, many researchers used hand-crafted techniques for the detection and classification of diseases. Also, due to a lack of knowledge experts and improper laboratory facilities, classification at an early stage is not possible. In terms of precision agriculture applications, developing effective, rapid, and highly successful computer-aided disease detection systems have become a necessity. A deep learning-based convolutional neural network (CNN) technique is used successfully in numerous computer vision tasks such as picture classification, object detection, segmentation, and image analysis for finding the best results. The existing models had been worked on a mixed crop approach and single class dataset with average accuracy. In this paper, the Modified InceptionResNet-V2 (MIR-V2) which is a form of the CNN model is used in conjunction with a transfer learning approach to recognize illnesses in images of tomato leaves. The proposed model is trained on a public dataset and self-collected dataset that includes seven different tomato leaf disease classifications with one healthy leaf. For the evaluation of model performance, different parameters such as dropout, learning rate, batch size, count of epochs, and accuracy are used. The disease classification accuracy rate for the applied models is 98.92% and the F1 score is 97.94%. The experimental results support the suggested approach’s validity and find it effective in detecting tomato leaf disease. © 2022, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Fast MIS on Incremental Graphs 31

Fast MIS on Incremental Graphs

引用

31st IEEE International Conference on High Performance Computing, Data, and Analytics Workshops, HiPCW 2024

作者： Trivedi, Aditya Nijhara, Prajjwal Banerjee, Dip Sankar Department of Computer Science and Engineering Indian Institute of Technology Jodhpur India

ISBN: (纸本)9798331509118

Maintaining the Maximal Independent Set (MIS) in dynamic graphs is crucial for applications like resource allocation, load balancing, and network optimization. Traditional approaches require complete re-computation, making them inefficient for large and evolving graphs. Our proposed methods utilize Tuple Valued Bitmap (TVB) and its extension with levels, TVBL, to enhance parallelism and minimize search space. Experimental evaluations on large-scale graphs, ranging from 50 million to 1.2 billion edges, demonstrate significant speedups-up to 15.64x for insertions, with minimal impact on the MIS cardinality. This approach offers a scalable solution for dynamic graphs, paving the way for further optimization in fully dynamic and streaming settings. © 2024 IEEE

关键词： Resource allocation

来源：评论

学校读者我要写书评

暂无评论

Tracking of Ship and Plane in Satellite Videos Using a Convolutional Regression Network with Deep Features 3rd

Tracking of Ship and Plane in Satellite Videos Using a Convo...

引用

3rd International Conference on Machine Vision and Augmented Intelligence, MAI 2023

作者： Sharma, Devendra Srivastava, Rajeev Computer Science and Engineering Department Indian Institute of Technology BHU Varanasi India

ISBN: (纸本)9789819743582

Object tracking is one of the latest research topic in various applications of computer vison. Satellite videos are provided by the commercial satellite makes it possible to extend into earth observation. Moving objects are occupy the small region of pixel and they are confused with neighboring regions also with objects. Due to the resolution constraints, the nearest similar objects can rarely be distinguished by appearance features. Conventional tracking techniques in satellite videos with hand-crafted features do not give better results. Deep learning-based techniques have proven to be superior in visual tracking benchmarks, but this is not fully explored for the object tracking in the satellite videos. Regression network is used to integrate the regression model with convolutional layers. Both motion feature and appearance feature are taken from pre-trained deep neural networks VGG16, VGG19, and ResNet50. When tracker faces obscure appearance information, motion features of object are used for discrimination and complementary information to improve the performances of model. Experimental analysis on ship and plane represents the ResNet50 feature extractor gives better tracking result. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Leveraging Convolutional Neural Networks for Enhanced Educational Question Classification

引用

SN computer science 2024年第8期5卷 1052页

作者： Deena, G. Ancy Breen, W. Raja, K. Shanmuganathan, C. Department of Computer Science and Engineering SRM Institute of Science and Technology Tamil Nadu Ramapuram Chennai India

Question classification is a critical task in natural language processing, as well as applications in question-answering systems and intelligent personal assistants. This study presents novel machine learning approach for automatic question classification based on Bloom’s taxonomy, a hierarchical structure of cognitive levels. Traditional methods rely on feature engineering and classic machine learning models, but these methods often lack the ability to understand complex patterns and connections in natural language data. This study proposes a novel approach using a convolutional neural network (CNN) model, which can automatically learn relevant features from input data without extensive feature engineering. The model classifies questions into pre-established categories or labels using an embedding layer, convolutional and pooling layers, and a fully linked layer for classification. We assess the model’s efficacy on a publicly accessible dataset generated by the subject expert and compare it to conventional machine learning methods like logistic regression and support vector machines. The CNN-based model outperforms conventional techniques, achieving greater recall, accuracy, precision, and F1-score. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Bloom’s Taxonomy Convolutional Neural Network Machine Learning Word embedding

来源：评论

学校读者我要写书评

暂无评论

Comparison of Tribological Behavior of Hard Faced Fe-Based Metallic Composite Coatings Sprayed Using HVOF

引用

Journal of Bio- and Tribo-Corrosion 2025年第1期11卷 1-18页

作者： Soni, Piyush Kumar Prasad, C. Durga Manjunath, H.R. Tiwari, Amit Raju, Vadali Pitchi Athithan, Senthil Deshmukh, Khemraj Kumar, Anjani Department of Mechanical Engineering New Horizon College of Engineering Karnataka Bengaluru India Department of Mechanical Engineering RV Institute of Technology and Management Karnataka Bengaluru India Department of Mechanical Engineering Ramaiah Institute of Technology Karnataka Bengaluru560054 India Department of Mechanical Engineering Suresh Gyan Vihar University Jaipur India Mechanical Engineering Department & amp Principal Indur Institute of Engineering and Technology Siddipet India Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Andhra Pradesh Vaddeswaram India Department of Biomedical Engineering Parul University Gujarat Vadodara India Mechanical Engineering Department Government Engineering College Arwal India

Fe-based alloys have recently become well-known for their ability to resist wear and corrosion. Iron, the primary component of alloys, is a cheaper option than alloys made from Co and Ni. There is a need for affordable and easily obtainable coatings based on iron. Coatings in the Fe–Ni–Cr system have demonstrated superior wear resistance compared to other commonly used Fe-based metallic coatings. The Fe-12Ni17Cr2.5Mo2.3Si0.03C based (Metco 41C) metallic coatings were applied onto boiler steel substrates using the High Velocity Oxy-Fuel (HVOF) spraying method. The coatings were examined using X-ray diffraction (XRD), a scanning electron microscope (SEM), and a microhardness tester to determine their density, porosity, microstructure, hardness, and surface roughness. The sliding wear behavior of the substrate and coatings was studied using the pin-on-disk tribometer. Figuring out the wear loss and coefficient of friction. Comparison was made of wear results between the substrate, coatings, and composite coatings. By combining these carbide powders with 30% SiC and WC-12Co particles, the coating’s wear resistance is enhanced in comparison to the Metco 41C coating. In comparison to boiler steel, the Metco 41C + SiC and Metco 41C + WC–Co coatings exhibited a roughly 70% and 40% reduction in volume loss compared with boiler steel respectively. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024.

关键词： HVOF thermal spraying

来源：评论

学校读者我要写书评

暂无评论

Effortless Meeting Analysis Automation 15

Effortless Meeting Analysis Automation

引用

15th International Conference on Advances in Computing, Control, and Telecommunication Technologies, ACT 2024

作者： Mohit, Harsh Chaitanya, K. Kumar, Maneesh Kushal, B. Shanthakumara, A.H. Department of Computer Science and Engineering Siddaganga Institute of Technology Tumakuru India

ISBN: (纸本)9798331300579

The "Automated DSM Insight Analyzer" is a groundbreaking initiative that revolutionizes the analysis of Daily Stand-Up Meetings (DSMs). Leveraging a sophisticated combination of cutting-edge technologies, including a Flask-based API for audio processing, Azure Speech-to-Text for precise transcription, and the GPT API for advanced Natural Language Processing (NLP), this solution provides comprehensive insights. It addresses challenges in resource management and bug tracking, ensuring the prompt generation of actionable insights. With multilingual support and seamless integration with Google Calendars, it enhances productivity, offering a user-friendly platform that fosters improved collaboration and project success through the extraction of valuable insights from daily meetings. © Grenze Scientific Society, 2024.

关键词： Application programming interfaces (API)

来源：评论

学校读者我要写书评

暂无评论

Prediction of Heart Disease Risk in Early Ages with Boosting Techniques 1

引用

1st International Conference on Machine Intelligence for Research and Innovations, MAiTRI 2023

作者： Yadav, Jyotsna Rahman, Habib Ur Department of Computer Science and Engineering Kanpur Institute of Technology Kanpur India

ISBN: (数字)9789819981359

ISBN: (纸本)9789819981342

Heart disease is one of the top causes of mortality globally. Early detection can save a person's life. In previous five years, it is found that young individuals are also experiencing heart disease. In order to automate the medical industry and to forecast cardiac illness, artificial intelligence (AI) provides a prediction model. We could employ a number of AI algorithms to predict the risk of heart disease as in previous work, done in this field;generally, authors have used single classifiers. In this work, we have used two approaches of AI. In the first approach, single or base classifiers like logistical regression (LR), support vector machine (SVM) and decision tree (DT) are examined. In the second approach, boosting classifiers such as adaptive boost (AdaBoost), extreme gradient boost (XGBoost) and gradient boost are analyzed, and then comparison is done between single classifiers and boosting classifiers and found that boosting model achieved highest accuracy and the most accurate model as compared with single classifiers in predicting heart disease risk in youngers. So a type of ensemble model called boosting model is employed to prevent overfitting and to offer the highest accuracy as compared to single models. With the XGBoost classifier, we obtained 96.29% accuracy, with AdaBoost classifier, we obtained 91.35% accuracy, and with the gradient boost classifier, we achieved the highest accuracy of 98.76%, which is best in prediction of heart disease ages between 28 and 50. © 2024, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Decision trees

来源：评论

学校读者我要写书评

暂无评论

MEC-EnergySaver: Unleashing Efficiency Through D2D and Data Compression 16

MEC-EnergySaver: Unleashing Efficiency Through D2D and Data ...

引用

16th International Conference on COMmunication Systems and NETworkS, COMSNETS 2024

作者： Ghosh, Anindita Mukherjee, Poulomi De, Tanmay National Institute of Technology Department of Computer Science and Engineering Durgapur India

ISBN: (纸本)9798350383119

The shift from basic cell phones to smartphones has ushered in intelligent devices, alongside a surge in mobile apps that reshape daily life. Simultaneously, technologies like data analytics, AI and IoT have enabled smart homes and advanced transport systems, demanding substantial computational power from mobile devices. To tackle this challenge, mobile-edge computing (MEC) has emerged as a solution. Unlike remote cloud computing, MEC offers computing services at the network's edge, allowing mobile users to transfer their computing tasks to MEC servers located nearby. However, MEC systems face challenges in efficient computation offloading and distributing resources in a way that reduces energy use and processing times. The evolution of mobile devices and the rise of data-intensive applications have led to the need for efficient computing solutions like MEC, with the potential to integrate data compression for energy savings. This paper introduced a two-phase approach for request servicing in MEC systems. The first phase prioritizes Device-to-Device (D2D) connections based on device proximity, efficiently pairing devices for data exchange. Unpaired devices are directed to the nearest MEC node. In the second phase, requests not fulfilled via D2D connections are handled by strategically positioned MEC nodes, ensuring comprehensive coverage and minimizing network energy consumption. © 2024 IEEE.

关键词： D2D communication data compression energy efficiency MEC

来源：评论

学校读者我要写书评

暂无评论

In-Group Bias in Deep Learning-Based Face Recognition Models Due to Ethnicity and Age

IEEE Transactions on Technology and Society

引用

IEEE Transactions on technology and Society 2023年第1期4卷 54-67页

作者： Nagpal, Shruti Singh, Maneet Singh, Richa Vatsa, Mayank Ratha, Nalini K. Ratha Indraprastha Institute of Information Technology Delhi Department of Computer Science and Engineering New Delhi110020 India Indian Institute of Technology Jodhpur Department of Computer Science and Engineering Jodhpur342011 India University at Buffalo Department of Computer Science and Engineering BuffaloNY14260 United States

Humans are known to favor other individuals who exist in similar groups as them, exhibiting biased behavior, which is termed as in-group bias. The groups could be formed on the basis of ethnicity, age, or even a favorite sports team. Taking cues from aforementioned observation, we inspect if deep learning networks also mimic this human behavior, and are affected by in-group and out-group biases. In this first of its kind research, the behavior of face recognition models is evaluated to understand if similar to humans, models also encode group-specific features for face recognition, along with where bias is encoded in these models. Analysis has been performed for two use-cases of bias: age and ethnicity in face recognition models. Thorough experimental evaluation leads us to several insights: (i) deep learning models focus on different facial regions for different ethnic groups and age groups, and (ii) large variation in face verification performance is also observed across different sub-groups for both known and our own trained deep networks. Based on the observations, a novel bias index is presented for evaluating a trained model's level of bias. We believe that a better understanding of how deep learning models work and encode bias, along with the proposed bias index would enable researchers to address the challenge of bias in AI, and develop more robust and fairer algorithms for mitigating bias as well as developing fairer models. © 2020 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Real-Time Caption Generation for the American Sign Language Using YOLO and LSTM

Real-Time Caption Generation for the American Sign Language ...

引用

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

作者： Jana, Urmi Paul, Subhashish Bhandari, Dinabhandhu Heritage Institute of Technology Department of Computer Science and Engineering Kolkata India

ISBN: (纸本)9798350382693

The lack of captions available for live video events on online platforms makes them inaccessible to the speech and hearing deprived communities. In this study, we examine two models: the pre-trained You Only Look Once (YOLO) framework and a Long Short Term Memory (LSTM) model that uses the Mediapipe framework from Google. The LSTM model is constructed by extracting the landmark features of the subjects face, torso and hands from the video frames and using them to predict the depicted gesture. This sets it apart from existing deep learning methods that use only a Convolutional Neural Network (CNN) framework to extract spatial features. We evaluate the performance and demonstrate the superiority of the proposed model against a LRCN model using our custom dataset. The proposed model offers a valuable approach to making live videos accessible to both viewers and speakers of the disabled community, who are currently depending on human translators. The ability to extract keypoints beforehand further enhances the models capabilities and speed. © 2024 IEEE.

关键词： LSTM Mediapipe Sign lanuage YOLO

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：