检索结果-内蒙古大学图书馆

SEDARU-net: a squeeze-excitation dilated based residual U-Net with attention mechanism for automatic melanoma lesion segmentation

引用

Multimedia Tools and applications 2024年 1-28页

作者： Lafraxo, Samira El Ansari, Mohamed Koutti, Lahcen Kerkaou, Zakaria Souaidi, Meryem LabSIV Department of Computer Science Faculty of Sciences Ibnou Zohr University Agadir80000 Morocco Department of Computer Sciences Faculty of Sciences Informatics and Applications Laboratory Moulay Ismail University Meknès Morocco

One of the most dangerous types of skin cancer, malignant melanoma, must be detected early on in order to receive successful therapy. If melanoma is not diagnosed in a timely manner, it might possibly result in death. According to clinical research, because of their varied hue, texture, and imperceptible borders, these early melanoma indications are very challenging for dermatologists to recognize. Therefore, it’s crucial to suggest an automated method that can accurately identify and distinguish between benign and malignant melanoma. Numerous automated methods have been developed by scientists to segment abnormalities from dermoscopic images. On the other hand, conventional models could find it difficult to reliably capture the multi-scale properties, which could result in inconsistent segmentation performance for a variety of object shapes and sizes. Furthermore, models with complicated forms and bounds, such as U-Net and DeepLabV3+, have difficulty properly segmenting tiny, thin, or complex lesions. Thus, we introduce A Squeeze-Excitation Dilated Residual U-Net with Attention Mechanism (SEDARU-Net) in this paper, a novel and automated semantic segmentation network for efficient skin lesion segmentation. To keep spatial information across layers and capture both local and global context, the model is built on U-net combined with dilated convolution. To solve optimization problems, residual blocks are used instead of the basic U-net units. This enhances feature learning, encourages better feature reuse, and allows for the creation of deeper and more robust networks. In order to encourage feature recalibration, global context awareness, and spatial adaptation, each residual block is supplemented with squeeze and excitation units. In addition, The attention gate is also included in the skip connection part of the network to enhance the beneficial channel dimension characteristics and suppress the unreliable background features. According to the results of the experim

关键词： Dermatology

来源：评论

学校读者我要写书评

暂无评论

Scaling Approach Over Learning Layer of Deep Learning Model to Reduce the FALSE Error in Binary Classification 8

Scaling Approach Over Learning Layer of Deep Learning Model ...

引用

8th International Conference on Electronics, Communication and Aerospace Technology, ICECA 2024

作者： Gokila, S. Helen, D. Alemu, Asmarech Muluneh Suresh, N.V. Hindustan Institute of Technology and Science Department of Computer Applications Chennai India Srm Institute of Science and Technology Department of Computer Applications Kattankulathur Chennai India Aset College of Science and Technology Chennai India

ISBN: (纸本)9798350367904

Objective: Worldwide Breast cancer is a main health issue for the women, every year thousands of deaths occurring due to late detection and diagnosis. computer-aided diagnostic detection systems are becoming a useful tool for early cancer cell identification. The proposed work uses the Deep Learning, Convolutional Neural Network (CNN), model to identify the cancerous cells in medical imaging accurately. Method: The proposed work balances the Learning Layers (LL) of CNN model to find the finest binary classification to detect whether the given image is Malignant or Benign. Experiment is carried out on gradual scaling of LL from 1 to 5 with learning cycles 10,15 and 20 respectively for each stage of scale to reduce the False Positive (FP), False Negative(FN) errors in each layer. Result: The performance of the scaled LL is evaluated using the metric like residual error loss, training accuracy in training stage. The model is evaluated with precision, recall, accuracy and f 1-score in the testing stage. By leveraging the power of deep learning, the system achieved an impressive accuracy rate of 97% in differentiating between the two common types of cancer cells in CNN-LL5. © 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

Breaking Barriers: An Innovative Algorithm to Stop Cyberbullying on Personal Blogs 4

Breaking Barriers: An Innovative Algorithm to Stop Cyberbull...

引用

4th International Conference on Emerging Systems and Intelligent Computing, ESIC 2024

作者： Anoop, M. Shyamala Devi, J. Srm Institute of Science and Technology Ramapuram Department of Computer Science and Applications Chennai India

ISBN: (纸本)9798350349856

The growing threat of cyberbullying appears as a serious and urgent worry in the modern digital landscape, where the internet serves as a canvas for personal expression through blogs and personal webpages. It is of the utmost significance to protect the valued freedom of internet artists and to uphold the purity of individual voices. We present a ground-breaking algorithm that was painstakingly designed to identify and stop cyberbullying in the context of personal blogs and webpages in response to this challenging challenge. Convolutional Neural Networks (CNNs), BERT (Bidirectional Encoder Representations from Transformers), and stemming approaches are all seamlessly combined in our ground-breaking strategy to create a robust defence against online abuse. Using CNNs' strong capabilities, it enables the detection of regional textual patterns and traits that could be used as red flags for cyberbullying. BERT goes deeply into the intricate tapestry of words, phrases, and sentences to ensure that the genuine intentions underlying the text are made clear thanks to its exceptional ability in contextual language interpretation. stemming procedure makes the detection system more precise and effective by assisting in the identification of dangerous words. © 2024 IEEE.

关键词： Blogs

来源：评论

学校读者我要写书评

暂无评论

Comparative Study of Machine Learning Algorithms towards Predictive Analytics

引用

Recent Advances in computer science and Communications 2023年第6期16卷 69-79页

作者： Petchiappan, Maheswari Aravindhen, Jaya Department of Computer Applications BS Abdur Rahman Crescent Institute of Science and Technology Vandalur Chennai India

Background: The trend of the stock market prediction has always been challenging and confusing for investors. There is tremendous growth in stock market prediction with the advancement of technology, machine learning, data science, and big data. The media and entertainment sector is one of the diverse sectors in the stock market. In the Indian stock market, Sensex and Nifty are the two indexes. The 2019 pandemic forced the movie theatres to shut down. As a result, distributors and film directors were not able to release their movies in theatres, and production was also stopped. Consequently, during the lockdown, people spent more time at home watching electronic media, resulting in a higher degree of media consumption. Objectives: The objective of the research is to predict the performance of the media and entertainment companies stock prices using machine-learning techniques. Investors will be benefited by maximizing the profit and minimizing the loss. Methods: The proposed stock prediction system is used to predict the stock values and find the accuracy of linear regression and logistic regression in machine learning algorithms for data science. Results: The experiments are conducted for the media and entertainment stock price data using Machine-learning algorithms. Media stock prices are considered as the input dataset. The model has been developed using the daily frequency of stock prices with different attributes. Conclusion: Thus, the media and entertainment stocks are predicted using linear regression and logistic regression. Using the above techniques, stock prices are predicted accurately to maximize profits and minimize the loss for the investors. © 2023 Bentham science Publishers.

关键词： Investments

来源：评论

学校读者我要写书评

暂无评论

Secure Online Banking Authentication System Using Time Bound Password 5

Secure Online Banking Authentication System Using Time Bound...

引用

5th IEEE International Conference on Computing, Power, and Communication Technologies, IC2PCT 2024

作者： Danthy, Ronson Pratham Pai, K. Rao, Vidya Manipal Institute of Technology Manipal Academy of Higher Education Department of Data Science and Computer Applications Manipal India Manipal Institute of Technology Manipal Academy of Higher Education Department of Computer Science and Computer Applications Manipal India

ISBN: (纸本)9798350383522

The banking system is a key component of financial transactions and economic expansion in the modern world. The introduction of internet banking, however, has created some brand-new difficulties, particularly in providing safe user authentication. This study investigates the flaws in conventional authentication techniques, including passwords and SMS-based One-Time Passwords (OTP), and suggests a creative fix. By using Time-Based One-Time Password (TOTP) created by an authenticator app and encrypting the user credentials with this as key, the proposed solution improves security. By addressing problems including poor password habits, SMS theft, OTP delivery delays, phishing, and social engineering attacks, this method strengthens the security of online banking. The approach provides a thorough framework for building this secure authentication system by detailing TOTP generation and encryption. This system successfully defends against popular attacks such as replay attacks, brute force attacks, rainbow tables, packet sniffing, and random guessing. © 2024 IEEE.

关键词： Advanced Encryption Standard - Galois Counter Mode (AES-GCM) Authentication Encryption Online Banking Time-Based One-Time Password (TOTP) Two-Factor authentication (2FA)

来源：评论

学校读者我要写书评

暂无评论

Predicting Risky Environment for Child Inside House using Deep Learning 5

Predicting Risky Environment for Child Inside House using De...

引用

5th International Conference on Emerging Smart Computing and Informatics, ESCI 2023

作者： Ahmad, Naeem Arya, Sumedha Singh, Deepak Department of Computer Applications Raipur India Department of Computer Science and Engineering Raipur India

ISBN: (纸本)9781665475242

We place a high priority on making sure kids are safe. Yet in the modern world, there are a lot of threats to children's safety and challenges to achieving other safety goals. Parent can't always supervise their kids and leave them alone inside the house. So this is a big problem of our time as the small kids is unable to assess the risk due to any harmful activity or objects. The purpose of this study is to provide safety to the child by alerting the parent about the child's position and surroundings. This proposed model can be used to monitor child's movements. We proposed a children's video data set. The video data set contains every moment of the child in the form of frames, and 30 frames are created every second. Additionally, we train a baseline computer vision module that determines the level of dangerousness for a child using deep learning. The baseline consists of well-established methods like conditional and logistic regression for child's pose estimation, as well as cutting-edge deep learning image classifiers. We then integrate the stumbled-condition baseline into various learning platforms and evaluate performance in terms of accuracy and loss. © 2023 IEEE.

关键词： computer vision

来源：评论

学校读者我要写书评

暂无评论

Enhanced Bayesian Optimized Support Vector Machine (BO-SVM) Classification and Prediction of Heart Disease 4

Enhanced Bayesian Optimized Support Vector Machine (BO-SVM) ...

引用

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

作者： Chairmadurai, P. Kavitha, P. Kamalakkannan, S. Pallavaram Department of Computer Science Chennai India Pallavaram Department of Computer Applications Chennai India Pallavaram Department of Information Technology Chennai India

ISBN: (纸本)9798331523923

Medical data analysis is increasingly used to plan, improve research techniques, and explain diagnoses. The allocation of medical resources is based on the prevalence of specific pathologies. Artificial intelligence (AI) offers robust software solutions to analyze existing data and provide optimal predictions of medical outcomes. The dataset for heart disease prediction consists missing values in important features like cholesterol, ECG results and lifestyle factors. This research focuses to solve the issues in the traditional machine learning algorithm and designing a system model incorporating multiple data processing algorithms to classify heart disease. The classification process allows for the generation of a predictive model from training and testing datasets. These datasets are processed using an algorithm for classification to create a new model that can analyze detailed data, potentially with these same classes, by combining computational techniques with mathematical tools. This study uses optimization strategies to improve this model's performance and maximize prediction outcomes. This study aims to provide a framework to forecast cardiac problems based on significant risk using various classifier techniques, including the Bayesian Optimized SVM, to variables. The UCI repository is the source of the dataset utilized in this investigation. Linear Discriminant Analysis is used for feature extraction to decrease dimensionality and enhance the model's capacity to identify relevant patterns in the data. To generate the results as a precision, Recall, f1 score and accuracy. © 2025 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

A Few-Shot Learning for HAR Using 3D Siamese Network 15

A Few-Shot Learning for HAR Using 3D Siamese Network

引用

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

作者： Guanghui, Hua Kumar, G. Hemantha Aradhya, V.N. Manjunath Punithavalli University of Mysore Dept. of studies in Computer Science Mysuru India Jss Science and Technology University Dept. of Computer Applications Mysuru India Bharathiar University Department of Computer Applications Coimbatore India

ISBN: (纸本)9798350370249

In artificial intelligence, deep learning techniques were applied in a wide field due to their ability to automatically learn complex patterns from raw data, such as image classification, object detection, image generation, and video analysis in computer vision. These applications are just the tip of a substantial iceberg. With the development of deep techniques to achieve state-of-art accuracy, learn precisely relevant features from raw data, and improve the robustness of models to make them versatile solutions. The traditional deep learning techniques require vast computational resources, including high-performance computer GPUs or TPUs, and can be expensive in terms of hardware. In Human Activity Recognition Systems (HARs). The video needs more computational resources than image processing. In this paper, we proposed a few-shot learning approach based on a 3D Siamese Network with depth video sequences to address the action video similarity or dissimilar problem. Convolutional Neural Networks (CNNs) are well-suited for images and spatial data processing. The 3DCNN is an extension of CNNs to incorporate an additional temporal dimension to capture the temporal pattern. From the 3D Siamese network, we extract more effective features with spatial-temporal action information from raw data. We employed the 3DCNNs Siamese Network (3DCNNs+SN) with shared weight to recognize or classify a new class of data without new sample data only based on learning training existing sample constructed model. In addition, we divided the benchmark dataset MSR Action 3D dataset into sub-dataset1 (SD-1), sub-dataset2 (SD-2) and sub-dataset3(SD-3) to train and validate our proposed method. For dissimilar sub-dataset 2 we got a 98.1% accuracy, 33% on high-similar sub-dataset 3 and 63.5% on merged sub-dataset SD-4 from sub-dataset2 (SD-2) and sub-dataset3(SD-3). © 2024 IEEE.

关键词： Spatio-temporal data

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Model for Parameter Tuning using SSO and CNN 3

A Hybrid Model for Parameter Tuning using SSO and CNN

引用

3rd International Conference on Intelligent Data Communication Technologies and Internet of Things, IDCIoT 2025

作者： Patidar, Anil Bandhu, Kailash Chandra Department of Computer Applications Medi-Caps University Indore India Department of Computer Science and Engineering Medi-Caps University Indore India

ISBN: (纸本)9798331527549

Social Spider Optimization (SSO) and Particle Swarm Optimization (PSO) algorithms applied to Intrusion Detection Systems (IDS) across small and large datasets. The study aims to evaluate and compare both PSO and SSO and find best optimal parameters in order to improve the accuracy of IDS. Additionally, the SSO algorithm find the best hyperparameter settings which are kernel, filter, padding, learning rate, and stride to achieve optimal results that directly impact IDS security. The evaluation of algorithms in terms of accuracy, precision, recall, F1 score, and AUC score. Results indicate that, on small datasets, PSO comes under overfitting but not SSO. Even on larger datasets, SSO crosses PSO significantly in terms of accuracy. This finding shows the importance of improvement in scalability, efficiency, and robustness of IDS with help of Social Spider Optimization (SSO). © 2025 IEEE.

关键词： Particle swarm optimization (PSO)

来源：评论

学校读者我要写书评

暂无评论

Creation of Cartoon Face Images of Celebrities: Automation 2nd

Creation of Cartoon Face Images of Celebrities: Automation

引用

2nd International Conference on Applied Algorithms, ICAA 2025

作者： Guru, D.S. Shivaprasad, D.L. Prajna, S. Department of Studies in Computer Science University of Mysore Manasagangotri Mysuru India Department of Computer Applications MIT First Grade College Mysuru India Department of Computer Applications JSS Science and Technology University Mysuru India

ISBN: (纸本)9783031845420

This paper addresses the problem of automatic detection and identification of celebrity cartoon faces and corpus creation. However, currently available face detection systems are mainly designed to detect human faces. We developed a celebrity car-toon face detector (CCFC) that specially detects celebrity cartoon faces featuring famous individuals, which is customised by a well-known object detector, Yolo-NAS, a highly efficient and lightweight deep architecture. Initially, we train our detector by employing a manually compiled dataset. Also, a suitable matching technique is utilised to create a celebrity cartoon face image database. The performance of the proposed system is evaluated with three different matching thresholds, and 90% accuracy is achieved in database creation. The novelty of the work lies in the complete setup of the process of creating a deployable database since this work is the first of its kind to automate the task of creating a corpus of images. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Automation Celebrity Cartoon Face Detector Database Creation

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：