检索结果-内蒙古大学图书馆

CACBL-Net: a lightweight skin cancer detection system for portable diagnostic devices using deep learning based channel attention and adaptive class balanced focal loss function

引用

Multimedia Tools and applications 2025年第14期84卷 12877-12900页

作者： Agrawal, Rishi Gupta, Neeraj Jalal, Anand Singh Department of Computer Engineering & Applications GLA University Mathura India

Accurate skin disease detection is one of the most challenging tasks due to high-class imbalance and limited labeled datasets. Recently Deep Convolutional Neural Network (DCNN) with ensemble learning has achieved significant popularity in skin cancer classification. However, implementing DCNN models with ensemble learning is not feasible for deployment on portable diagnostic devices due to the limitation in computing resources and computing time. This paper proposes a Channel Attention and Adaptive Class Balanced Focal Loss function based lightweight Deep CNN model (CACBL-Net) for handling the issues of data imbalance and limited computing resources of portable diagnostic devices, such as mobile phones or tablets. Channel attention explores interdependencies between channels by recalibrating channel-wise feature responses. To deal with the issue of high-class imbalance, the proposed method used an adaptive class balance focal loss function which can quickly concentrate the model on complex cases while automatically downweighting the contribution of easy examples during training. The proposed CACBL-Net is validated on three popular skin cancer datasets which are HAM-10000, PAD-UFES-20, and MED-NODE. Dermoscopic, non-dermoscopic and smartphone images are taken from all three datasets for experimental work. The quantitative findings indicate that the proposed CACBL-Net model achieved a sensitivity of 90.60%, 91.88%, and 91.31% for the HAM-10000, PAD-UFES-20, and MED-NODE datasets, respectively. Additionally, the average prediction time per patient was recorded at 0.006, 0.010, and 0.011 s. These results demonstrate superior performance compared to other state-of-the-art deep learning models. The experimental finding suggested that the proposed method can achieve a significant performance at a low cost of computational resources and inference time, which makes it potentially feasible for deployment in portable diagnostic devices for automated diagnosis of skin lesions.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Physiological biometric image quality assessment -a review

引用

Multimedia Tools and applications 2025年 1-35页

作者： Choudhary, Parul Pathak, Pooja Gupta, Phalguni Department of Computer Engineering & Applications GLA University U.P Mathura India

This paper makes a literature survey on biometric image quality assessment (BIQA) techniques focusing on physiological traits. It covers a wide range of methodologies, metrics and evaluation techniques. Objective image quality assessment (IQA) methods primarily focus on quantifying image quality using computational algorithms, while subjective IQA methods rely on human observers to provide quality ratings based on visual perception. The paper categorises the existing IQA techniques based on their characteristics and applications. It explores different types of distortion models, such as noise, blur, compression artefacts, and colour inconsistencies that are commonly encountered in digital images. Additionally, it investigates the influence of various factors on image quality, including image content, context and viewer preferences. The survey concludes by summarising the key findings and identifying the current trends and future directions in BIQA research. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.

关键词： Biometrics

来源：评论

学校读者我要写书评

暂无评论

Hybrid deep learning and machine learning approach for detecting spatial and temporal forgeries in videos

引用

Neural Computing and applications 2024年 1-15页

作者： Singh, Upasana Rathor, Sandeep Kumar, Manoj Department of Computer Engineering & amp Applications GLA University Uttar Pradesh Mathura India

Video forgery detection has become increasingly critical due to the rise of sophisticated video manipulation techniques. Traditional methods often struggle to keep up with the ever-evolving sophistication of forgery techniques, necessitating innovative and advanced solutions. In response to this challenge, we propose a hybrid approach that combines deep learning and machine learning techniques. This approach leverages the power of DenseNet121 for excellent spatial feature extraction, utilizes a spatiotemporal autoencoder to capture spatiotemporal dependencies, and employs a gradient boosting machine (GBM) to effectively classify the video clips based on the extracted features. This provides a robust and comprehensive solution for detecting spatial and temporal forgeries in videos. By combining these components, this approach offers a more robust and comprehensive solution for detecting spatial and temporal forgeries in videos. To evaluate our approach, we are using a diverse dataset of videos (VTD and VIFFD) with different forgery scenarios and conducting extensive experiments. The results demonstrate the effectiveness and accuracy of our hybrid approach in detecting spatial and temporal forgeries, surpassing the performance of individual classifiers and traditional methods. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Growth model for detection and removal of faults having different severity with single change point and imperfect debugging

引用

International Journal of Reliability and Safety 2024年第1期18卷 86-101页

作者： Tiwari, Asheesh Sharma, Ashish Department of Computer Engineering & Applications GLA University Mathura Uttar Pradesh Mathura India

Throughout the last decades, researchers have modelled a variety of software reliability growth models for estimating measures of reliability. In the present paper, we have classified faults into four divergent types as per their easiness and hardness in detection and removal. Also, variations in fault detection and correction rates can be because of the testing strategy, changing testing environment, motivation, proficiency and organisation of the debugging and testing teams, etc. In the present paper, a change point has been applied to four types of faults along with imperfect debugging during the correction of faults. This paper comprises two proposed software reliability growth models, which are compared on the basis of rate of detection as well as correction. All the model parameters are evaluated by the method of least squares. These models are assessed using various comparison measures like SSE, MSE, RMSE, Bias, variance and RMSPE. Copyright © 2024 Inderscience Enterprises Ltd.

关键词： Software reliability

来源：评论

学校读者我要写书评

暂无评论

Automatic Classification of Melanoma Using Grab-Cut Segmentation & Convolutional Neural Network

引用

SN computer Science 2024年第5期5卷 591页

作者： Verma, Sarvachan Kumar, Manoj Computer Engineering & amp Applications GLA University Mathura India

Background: Skin diseases are common health complications around the world. One of the most unsafe types of skin cancer is melanoma. Detection of skin cancer in the early stage can reduce the mortality rate. Manually detecting skin cancer from dermoscopy images is time-consuming, difficult, expensive, and requires experience. Therefore computer-based systems are necessary to make an early finding of skin cancer to plan timely treatment for patients to increase their survival rates. In this work, we introduce a Gaussian filter for noise removal. Furthermore, we applied the Grabcut segmentation technique to segment the affected lesions. Then we applied the segmented image into Convolutional Neural Network (CNN) with three hidden layers, including a flat and dense layer. The proposed CNN model uses the Adam optimizer to deliver the best performance with an accuracy value of 97.50%. In this study, we used 3500 images containing 500 images of each class. The data is further split into training and validation data as 75% and 25%. Based on the proposed model described in Fig. 2, segmented images are of size (32,32,3) as an input to the CNN model, which contains three hidden layers. The idea is convoluted, using 3 × 3 filters of 256,1n28 and 64, respectively, on each hidden layer. This Paper used the Convolutional Neural Network based method proposed for seven types of melanoma classification. It aids Skin cancer classifications that should be detectable by both patients and clinicians. By the experimental and evaluation, it can be said the designed model can be considered as a benchmark for skin cancer classification. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Convolutional neural network Dermoscopy Image Classification Melanoma Skin Cancer

来源：评论

学校读者我要写书评

暂无评论

Sentinel Exploring the Effectiveness of Autonomous IoT Based Security Robots in Enhancing Surveillance 1

Sentinel Exploring the Effectiveness of Autonomous IoT Based...

引用

1st International Conference on Advanced Computing and Emerging Technologies, ACET 2024

作者： Kumar, Ajitesh Kumari, Mona Computer Engineering and Applications Gla University Mathura India

ISBN: (纸本)9798350367720

This paper introduces an IIoT driven intelligent Security based robot "SENTINEL", leveraging an Arduino Uno, camera unit, ultrasonic sensor, motor controller, motors,The envisioned robot is engineered to autonomously survey a specified area, capturing images and videos with the camera unit. The ultrasonic sensor identifies obstacles to avert collisions, alerting the user. The robot maneuvers and alters its course through the motor controller and motors, managed by an Arduino Uno. This system finds utility in various domains, including surveillance and security, enhancing security-based system. It is economically developed, broadening its accessibility for future used. The system's deployment has undergone testing, demonstrating proficiency in recognizing and responding to environmental cues. A web- based interface facilitates remote control and monitoring by users. © 2024 IEEE.

关键词： Ultrasonic sensors

来源：评论

学校读者我要写书评

暂无评论

CNN Image Forgery Detection: Machine Learning Improves Photojournalism Accuracy and Robustness 3

CNN Image Forgery Detection: Machine Learning Improves Photo...

引用

3rd International Conference on Communication, Control, and Intelligent Systems, CCIS 2024

作者： Thakur, Varsha Agarwal, Rohit Gla University Department of Computer Engineering & Applications Mathura India

ISBN: (纸本)9798331528201

The purpose of this study is to address the rising problem of picture modification and fraud, especially in the field of photojournalism, where the ability to recognize changed photos is essential to preserving faith in digital technology. By employing CNN-based approaches, the created forgery detection system exhibits its capacity to reliably discern between actual and AI-generated photos, reaching a balanced classification of 49% and 50%, respectively, with low misclassification. The purpose of this research is to investigate the development of digital picture forensics by using both conventional methods and contemporary innovations such as blockchain and deep learning. In order to achieve a higher level of detection accuracy, the suggested system makes use of machine learning models, such as transfer learning using pretrained models like ImageNet. The findings demonstrate that the system is capable of generalization and has a limited amount of overfitting, as seen by the high training accuracy of 98.44% and the validation accuracy of 98.92%. It is accepted that there is a need for future improvements in scalability, robustness to adversarial assaults, and interpretability, despite the fact that the findings have been somewhat encouraging. In order to successfully battle increasingly complex picture forgeries, future breakthroughs in digital image forensics will need cooperation between the government, industry, and academic institutions. © 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

A Deep Learning Based Strategies for Scene-Text VQA System used in Industrial applications: A Critical Analysis 1

A Deep Learning Based Strategies for Scene-Text VQA System u...

引用

1st International Conference on Innovative Sustainable Technologies for Energy, Mechatronics and Smart Systems, ISTEMS 2024

作者： Agrawal, Mayank Singh Jalal, Anand Sharma, Himanshu Gla University Department of Computer Engineering and Applications Mathura India

ISBN: (纸本)9798350384246

Scene-Text Visual Question Answering (Scene-Text VQA) is an emerging research area that combines computer vision, natural language processing, and scene understanding. The goal of Scene-Text VQA is to develop models and algorithms that can comprehend scene text and accurately answer questions based on textual and visual information present in images. Scene-Text VQA models used in different areas such as quality control and inspection, equipment maintenance, safety compliance, supply chain visibility, industrial automation, industry 4.0, assistance for the visually impaired, document analysis, autonomous vehicles, e-commerce, smart cities, cultural heritage preservation, surveillance, and security. Scene-Text VQA models face unique challenges, including accurate text detection and recognition, understanding contextual cues, handling language complexities, fusing text and image features effectively, limited training data, biases in datasets, generalization to unseen text and scenes, computational complexity, and appropriate evaluation metrics. In this paper we will explain different methods and algorithms for Scene-Text VQA models that can comprehend scene text and accurately answer questions based on textual and visual information present in images. The authors will distinguish all the techniques of Scene-Text VQA for effectively process and comprehend the textual information present in images and show how can one technique is conflicting then the other as well. © 2024 IEEE.

关键词： Inspection

来源：评论

学校读者我要写书评

暂无评论

Optimizing SMS Spam Detection: An In-Depth Analysis of Machine Learning Approaches 5

Optimizing SMS Spam Detection: An In-Depth Analysis of Machi...

引用

5th International Conference on Data Intelligence and Cognitive Informatics, ICDICI 2024

作者： Jain, Vinod Gla University Department of Computer Engineering and Applications Mathura India

ISBN: (纸本)9798350389609

Since mobile messaging has become more popular, there is a rising worry in the digital communication environment about SMS spam detection because of the surge in unwanted and potentially hazardous communications. Beyond conventional rule-based approaches, machine learning (ML) provides a potent way to automate the detection and filtering of spam in SMS (Short Message Service) systems. By converting text data into numerical representations, these techniques enable models to identify trends in both spam and ham (non-spam) communications. SMS spam datasets are used to train the models, and measures including precision, recall, and F1-score are used to assess the *** outcomes show that machine learning models are capable of identifying spam. Because they work well in text classification tasks, Naive Bayes and Logistic Regression in particular perform well. According to the study's findings, machine learning provides a reliable, scalable, and adaptable method for detecting SMS spam that can handle big datasets and changing spam strategies. Advanced NLP (Natural Language Processing) techniques is integrated into SMS systems to greatly enhance the performance of ML models. Future research will include multilingual spam filtering and real-time detection. © 2024 IEEE.

关键词： Logistic regression

来源：评论

学校读者我要写书评

暂无评论

Enhancement of Logistic Regression to Detect Breast Cancer using Cross Validation 8

Enhancement of Logistic Regression to Detect Breast Cancer u...

引用

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

作者： Jain, Vinod Gla University Department of Computer Engineering and Applications Mathura India

ISBN: (纸本)9798350367904

Many women in the world are suffering from cancer. Breast cancer is among one of the cancer which is affecting women all over the globe. The machine learning algorithms are found good for detecting breast cancers. Logistic Regression is among the best supervised machine learning algorithms which is found very good in detecting breast cancer. Many hospitals provided their data of females which comes for breast cancer detection. WDBC is one of the best known data sets for breast cancer. K-fold cross-validation reduces the possibility of overfitting and enhances generalization by repeating the process across all folds and computing the average performance. By applying the Logistic Regression model with cross validation we can more effectively generalize to new data. Through the identification of the most significant predictors, feature selection further enhances the model by lowering computational complexity and dimensionality. The findings show that good accuracy for breast cancer prediction is obtained when using Logistic Regression in conjunction with cross-validation. The model is a good fit for clinical applications because of its robust performance, ease of interpretation, and simplicity. The findings of this study demonstrate the validity of logistic regression as a tool for the diagnosis of breast cancer and imply that its combination with cross validation improves diagnostic accuracy and aid in clinical decision-making. © 2024 IEEE.

关键词： Logistic regression

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：