检索结果-内蒙古大学图书馆

MOPO-HBT: A movie poster dataset for title extraction and recognition

Multimedia Tools and Applications 2024年第18期83卷 54545-54568页

作者： Ghosh, Mridul Roy, Sayan Saha Banik, Bivan Mukherjee, Himadri Obaidullah, Sk Md Roy, Kaushik Dept. of Computer Science Shyampur Siddheswari Mahavidyalaya Howrah India Dept. of Computer Science & Engineering Aliah University Kolkata India Dept. of Electronics & Electrical Communication Engineering IIT Kharagpur Kharagpur India TISA Lab Dept. of Computer Science West Bengal State University Barasat India

Real-world images often encompass embedded texts that adhere to disparate disciplines like business, education, and amusement, to name a few. Such images are graphically rich in terms of font attributes, color distribution, foreground-background similarity, and component organization. This aggravates the difficulty of recognizing texts from these images. Such characteristics are very prominent in the case of movie posters. One of the first pieces of information on movie posters is the title. Automatic recognition of movie titles from images can aid in efficient indexing as well as information conveyance. However, it is accompanied by other texts like the names of actors, producers, taglines, dates, etc. Though the organization of components is somewhat similar across different film industries like Tollywood (West Bengal), Bollywood (Mumbai), and Hollywood (Los Angeles), the graffiti patterns differ in multifarious instances. To address the problem of movie title understanding, we propose a dataset named MOvie POsters-Hollywood Bollywood Tollywood (MOPO-HBT) that encompasses movie posters from the aforementioned industries. The entire dataset is publicly available (http://***/11564) for research purposes. The baseline results of title identification and recognition were obtained with a CNN-based (Convolutional Neural Network) approach, wherein the titles were extracted using the M-EAST (Modified-Efficient and Accurate Scene Text) detector model. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2023.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

Water Body Detection Utilizing NDWI, NDVI and NMDWI Indices in SEN-12 Spectral Imagery 1

Water Body Detection Utilizing NDWI, NDVI and NMDWI Indices ...

引用

1st International Conference on electronics, communication and Signal Processing, ICECSP 2024

作者： Vasanthi, A. Joshitha, K. Lakshmi Sri Sairam Engineering College Department of Computer Science and Engineering Chennai India Sri Sairam Engineering College Department of Electronics & Communication Engineering Chennai India

ISBN: (纸本)9798350364590

This paper investigates the efficacy of water body detection utilizing three widely used spectral indices, namely Normalized Difference Water Index (NDWI), Normalized Difference Vegetation Index (NDVI), and Modified Normalized Difference Water Index (NDWI), in SEN-12 spectral imagery. These indices are computed from the diverse spectral bands available in Sentinel-2 data to enhance our ability to identify and analyze water bodies. The SEN-12 dataset provides a rich source of multi-spectral information, enabling a nuanced exploration of these indices for water detection. Methodologies involve calculating and comparing NDWI, NDVI, and MNDWI across various scenes, assessing their strengths and limitations in different environmental contexts. The findings contribute valuable insights into the optimization of spectral indices for water body detection, thereby enhancing the utilization of remote sensing data for hydrological applications and environmental monitoring. © 2024 IEEE.

关键词： Environmental monitoring

来源：评论

学校读者我要写书评

暂无评论

Implementing a Robust Method for Detecting Human Actions in Health Monitoring by Employing Sensors on Mobile Internet of Things Devices and Utilizing a One-Dimensional Convolutional Neural Network

引用

Journal of Technology in Behavioral science 2025年第1期10卷 79-91页

作者： Vignesh, T. Thyagharajan, K.K. Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Andhra Pradesh Vaddeswaram India Department of Electronics and Communication Engineering R.M.D Engineering College Chennai India

This research is focused on developing a precise human action recognition system by utilizing accelerometer and gyroscope sensor data. The aim is to analyze various human actions like jogging, sitting, standing, walking, going upstairs, going downstairs, and falling, using multiple sensors found in modern smartphones. Previous studies have explored human action recognition using different sensors, cameras, and machine learning and deep learning techniques. However, some existing approaches, particularly those based on channel state information (CSI), have shown less satisfactory outcomes. In contrast, this paper employs deep learning techniques, specifically one-dimensional convolutional neural networks (1D-CNN), for more accurate human action analysis. One significant advantage of the proposed system is its potential as a wearable device. The paper will introduce a one-dimensional convolutional neural network tailored for human action recognition and an optimized approach for determining the magnitude of accelerometer's X, Y, and Z coordinates, as well as gyroscope sensor values. By utilizing these coordinates, the One-Dimensional Convolutional Neural Networks can effectively classify various human actions like jogging, sitting, standing, walking, going upstairs, going downstairs, and falling. To evaluate the performance of the 1D-CNN approach, a comparison is made with other methods, including SOM, SVM, and FCM. The results demonstrate that the proposed method (1D-CNN) outperforms other classification techniques with an impressive accuracy of 92.42 per cent. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024.

关键词： Acceleration Gyroscope Human action recognition One-dimensional convolutional neural network Sensors

来源：评论

学校读者我要写书评

暂无评论

Next-Gen SDN Security: Quantum Keying Meets AI Analytics

Next-Gen SDN Security: Quantum Keying Meets AI Analytics

引用

2024 International Conference on Distributed Systems, computer Networks and Cybersecurity, ICDSCNC 2024

作者： Virushabadoss, S. Anithaashri, T.P. Nivedha, M. Krishnan, S. Hari Karuppannan, Anand Saveetha Institute of Medical and Technical Sciences Saveetha School of Engineering Dept. of Computer Science and Engineering Chennai India Arasu Engineering College Department of Electronics and Communication and Engineering Kumbakonam India Sanskrithi School of Engineering Department of Electronics and Communication and Engineering Puttaparthi India Gnanamani College of Technology Department of Electronics and Communication and Engineering Namakkal India

ISBN: (纸本)9798350375442

This paper suggests a complete security architecture that blends quantum key distribution (QKD) with cutting-edge machine learning methods to improve the Software-Defined Networking (SDN) control plane. By creating secure channels of communication through QKD, the framework ensures integrity and secrecy of control messages between the SDN controller and network devices. Furthermore, network traffic is continuously monitored and analyzed using machine learning techniques, enabling the detection of unusual behaviors that might be warning signs of upcoming security problems. Tests utilizing Mininet as the network interface and Open Daylight as the SDN controller proved the efficacy of the suggested method. The outcomes revealed a minimal computational overhead, exceptional detection accuracy up to 98.1%, and low false positive rates, substantiating the system's ability to effectively manage varied network conditions and traffic volumes. These results underscore the potential of integrating QKD with robust machine learning strategies to secure the SDN control plane against sophisticated threats. © 2024 IEEE.

关键词： Quantum cryptography

来源：评论

学校读者我要写书评

暂无评论

A video data hiding technique based on pixel sequence, weight interpolation and quorum function

引用

Multimedia Tools and Applications 2024年第41期83卷 88881-88901页

作者： Roselinkiruba, R. Jothi, C. Saranya Koushik, C.P. Keerthika, A. Sathya, Vanlin School of Computer and Science Engineering Vel Tech Rangarajan Dr.Sagunthala R & D Institute of Science and Technology Chennai Avadi India School of Computer and Science Engineering VIT Bhopal University Bhopal India Electronics and Communication Engineering Aarupadai Veedu Institute of Technology Chennai India Celona Inc CampbellCA United States

The development of multimedia technology has increased the challenge in protecting the information. This paper proposes an effective method for data hiding in videos based on the Pixel Sequence (PS), weight interpolation to protect the secret data from intruders. Initially, the noise is removed from the video with the help of median filter and Moving Objects (MO) inside the frames having different backgrounds are accurately detected using Fast Region-based Convolution Neural Network (FR-CNN). Next, the MO and Non Moving Object (NMO) pixel from each frame are mapped with the adjacent frames to determine the weights applying PS. Then, the cover video is up-scaled with interpolated pixels based on weights of MO and NMO. Finally, the secret data are hidden inside the both interpolated and non-interpolated pixels elevate Embedding Capacity (EC). In interpolated pixels, both Most Significant Bit (MSB) and Least Significant Bit (LSB) of each frames based on their weight are utilized to hide data. Whereas, in non-interpolated pixels, Quorum Function (QF) is used to embed data. Two different data embedding techniques are adopted in each frame to enhance the security. The performance evaluation of the proposed method out-performs the existing methods in terms of precision and recall. Experimental results also show that the stego-video quality is uncompromised and is resistive to security attacks. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of ML Algorithms for Forecasting Time Series Stock Returns 4

Comparative Analysis of ML Algorithms for Forecasting Time S...

引用

4th International Conference on Sentiment Analysis and Deep Learning, ICSADL 2025

作者： Bondili, Bala Karthikeya Singh Bolagani, Sai Pavan Thatikonda, Rakshith Vnr Vignana Jyothi Institute of Engineering and Technology Computer Science and Engineering Telangana India Koneru Lakshmaiah Education Foundation Computer Science and Engineering Andra Pradesh India Cmr Technical Campus Electronics and Communication Engineering Telangana India

ISBN: (纸本)9798331523923

Forecasting stock returns is a challenging task due to high data variability, noise, and the inherent unpredictability of financial markets. Traditional methods like ARIMA and regression lack scalability and fail to capture complex data patterns. Recent advancements in machine learning, including ensemble methods such as Random Forest and Boosting, have shown promise in improving predictive accuracy. However, these methods face challenges such as overfitting, computational complexity, and data dependency. This paper evaluates the performance of various machine learning algorithms and their ability to address these challenges to improve forecast reliability. By using techniques like dimensionality reduction and hyperparameter tuning, the study identifies strategies to enhance prediction accuracy while reducing computational complexity and costs. © 2025 IEEE.

关键词： Dimensionality reduction

来源：评论

学校读者我要写书评

暂无评论

Secure electronics Medical Infrastructure for Healthcare 4.0: A Voice Identity Management-Based Approach 2

Secure Electronics Medical Infrastructure for Healthcare 4.0...

引用

2nd International Conference on Machine Learning and Data engineering, ICMLDE 2023

作者： Pal, Prashnatita Sahana, Bikash Chandra Poray, Jayanta Department of Electronics & Communication Engineering National Institute of Technology Patna India Department of Electronics & Communication Engineering National Institute of Technology Patna India Department of Computer Science & Engineering Techno India University Kolkata India Department of Electronics & Communication Engineering St Thomas College of Engineering & Technology Kolkata India

Securing Electronic Medical Records for Healthcare 4.0 is one of the prime requirements in the medical care service domains. Healthcare 4.0 refers to the evolution process of healthcare systems that leverage digital technologies, connectivity, and data-driven insights to provide more personalized and efficient care. With the increasing adoption of electronic healthcare records and the growing importance of data in the healthcare sector, ensuring the privacy of patient data becomes paramount. To ensuring the security for healthcare 4.0 in the current work a voice authentication-based system is devised. This enables the protection of electronic healthcare records (EHRs) in the association of healthcare 4.0. More precisely here the voice authentication is used as an additional layer of security for accessing and managing EHRs. The Voice authentication can also be categorized as the voice biometric. This is a technology that uses unique voice characteristics and patterns to verify the identity of individuals. The result establishes the innovative smart security arrangement for the Healthcare 4.0 using the voice authentication-based identity management infrastructure;and a modified version of asymmetric cryptosystem has been used inside the blockchain. The simulation outcome indicates better performance in the form of block generation time, block number, ensure the transaction, and minimize the latency of the transaction. © 2024 Elsevier B.V.. All rights reserved.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

Intelligent traffic prediction system using hybrid convolutional neural networks for smart cities

引用

Multimedia Tools and Applications 2024年 1-19页

作者： J, Jeba Sonia G, Arun Kumar kumar, ERajesh Raju, K. Narasimha Sudha, V. Kshirsagar, Pravin R. Tirth, Vineet Rajaram, A. Department of Data Science and Business Systems School of Computing College of Engineering and Technology SRM Institute of Science and Technology Tamil Nadu Chennai603203 India Department of Electronics and Communication Engineering JSS Academy of Technical Education Uttar Pradesh Noida201301 India Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Andhra Pradesh Vaddeswaram522302 India Andhra Pradesh Visakhapatnam530048 India Department of Electronics and Communication Engineering Sona College of Technology Tamil Nadu Salem636005 India Department of Electronics & amp Telecommunication Engineering J D College of Engineering & ampManagement Maharashtra Nagpur India Mechanical Engineering Department Centre for Engineering and Technology Innovations College of Engineering King Khalid University Asir Abha61421 Saudi Arabia Department of Electronics and Communication Engineering EGS Pillay Engineering College Nagapattinam India

Intelligent Traffic Management Systems (ITMS) play a crucial role in advancing smart city applications by addressing the rapid increase in road traffic due to the growing number of vehicles. Effective real-time traffic prediction is essential to avoid congestion, yet existing systems often struggle with high delays and insufficient accuracy. This paper introduces an innovative ITMS based on a Hybrid Convolutional Neural Network (HCNN) to overcome these challenges. The proposed system integrates numerous smart devices, including smart vehicles, with an edge computing framework where the HCNN model is deployed. The HCNN combines Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to enhance prediction accuracy and reduce latency. Experimental results demonstrate that the HCNN model achieves up to 95% prediction accuracy while reducing delays by up to 50% compared to conventional methods. These improvements are significant for smart city applications, ensuring timely and precise traffic management. The study validates the effectiveness of the HCNN model in improving traffic prediction and proposes its potential for broader smart city implementation. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

An Adaptive Machine Learning Approach for Electrical Fault Detection and Minimization 2

An Adaptive Machine Learning Approach for Electrical Fault D...

引用

2nd International Conference on Information and communication Technology, ICICT 2024

作者： Biswas, Md. Emran Ahamed, Md Sayem Rafi, Md. Abid Hasan Haque, Md Dulal Department of Electronics and Communication Engineering HSTU Bangladesh Department of Computer Science and Engineering Engineering UIU Bangladesh Shanghai Institute of Technical Physics Chinese Academy of Science China

ISBN: (纸本)9798331508227

In the field of electrical power systems, effective fault detection, and minimization are crucial for maintaining system reliability and operational efficiency. This paper introduces machine learning based Adaptive Fault Reduction Algorithm (AFRA), a novel performance optimization approach designed to reduce fault occurrences significantly of electrical power systems. We used a binary classification dataset with a total of 12,001 data samples for this research. Random Forest (RF) model was used for fault detection and compared its performance against the Artificial Neural Network (ANN) model. RF model demonstrated accuracy, precision, and recall, with scores of 99.79%, 99.64%, and 99.91% respectively, compared to the ANN’s 99.54%, 99.54%, and 99.45%. By integrating AFRA, we reduced the number of fault events from 5,496 to just 6, achieving a 99.89% reduction. Our approach highlights the importance of optimization in fault detection and minimization methodologies, offering a robust framework for improving system reliability. The promising results achieved by AFRA suggest significant potential for its application in real-time monitoring and dynamic fault detection and minimization in power systems. ©2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Multimodal and Multi-Temporal in Satellite Images Based on Flood Detection: A Comparative Study 2

Multimodal and Multi-Temporal in Satellite Images Based on F...

引用

2nd IEEE International Conference on Data science and Network Security, ICDSNS 2024

作者： Sonavale, Amruta Chakkaravarthy, Midhun Rao, Surampudi Srinivasa Salleh, Hishamuddin Bin M. Jadhav, Jagannath Lincoln University College Department of Computer Science and Multimedia Malaysia Malla Reddy College of Engineering and Technology Department of Electronics and Communication Engineering Telangana India Kle College of Engineering and Technology Department of Electronics and Communication Engineering Chikodi India

ISBN: (纸本)9798350373110

Floods can cause significant damage to human lives and property, necessitating precise post-flood evaluations for rescue efforts. Methods: It explores flood location AI techniques using a multi-modular and multi-fleeting picture dataset, ranging from traditional methods like multi-facet perceptron (MLP), support vector machine (SVM) and Deep convolution neural network (DCNN) to space transformation-based strategies. We utilized SPOT-5 and RADAR pictures from the November 2000 flood in Gloucester, UK. Results: The F1 score of 0.8846 and the region under the accuracy (AUC) bend (PR) of 0.9173 in the examinations were marginally higher than those that used machine strategies and investment of human resources. Conclusion: Utilizing a multi-modular and multi-temporal image dataset looked at a few customary flood location AI techniques as MLP, SVM and DCNN, to later space transformation-based strategies. This enhancement is due to the 20 labelled data samples used in the domain adaptation- based on semi-supervised domain adaptation (SSDA). © 2024 IEEE.

关键词： Investments

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：