Pub Date : 2023-08-16DOI: 10.3991/ijoe.v19i11.40267
Anmar A. Al-Janabi, Sufyan T. Faraj Al-Janabi, Belal Al-Khateeb
Deep learning and its variant techniques have surpassed classical machine algorithms due to their high performance gaining remarkable results and are used in a broad range of applications. However, adopting deep learning models over the cloud introduces privacy and security issues for data owners and model owners, including computational inefficiency, expansion in ciphertext, error accumulation, security and usability trade-offs, and deep learning model attacks. With homomorphic encryption, computations on encrypted data can be performed without disclosing its content. This research examines the basic concepts of homomorphic encryption limitations, benefits, weaknesses, possible applications, and development tools concentrating on neural networks. Additionally, we looked at systems that integrate neural networks with homomorphic encryption in order to maintain privacy. Furthermore, we classify modifications made on neural network models and architectures that make them computable via homomorphic encryption and the effect of these changes on performance. This paper introduces a thorough review focusing on the privacy of homomorphic cryptosystems targeting neural network models and identifies existing solutions, analyzes potential weaknesses, and makes recommendations for further research.
{"title":"Secure Data Computation Using Deep Learning and Homomorphic Encryption: A Survey","authors":"Anmar A. Al-Janabi, Sufyan T. Faraj Al-Janabi, Belal Al-Khateeb","doi":"10.3991/ijoe.v19i11.40267","DOIUrl":"https://doi.org/10.3991/ijoe.v19i11.40267","url":null,"abstract":"Deep learning and its variant techniques have surpassed classical machine algorithms due to their high performance gaining remarkable results and are used in a broad range of applications. However, adopting deep learning models over the cloud introduces privacy and security issues for data owners and model owners, including computational inefficiency, expansion in ciphertext, error accumulation, security and usability trade-offs, and deep learning model attacks. With homomorphic encryption, computations on encrypted data can be performed without disclosing its content. This research examines the basic concepts of homomorphic encryption limitations, benefits, weaknesses, possible applications, and development tools concentrating on neural networks. Additionally, we looked at systems that integrate neural networks with homomorphic encryption in order to maintain privacy. Furthermore, we classify modifications made on neural network models and architectures that make them computable via homomorphic encryption and the effect of these changes on performance. This paper introduces a thorough review focusing on the privacy of homomorphic cryptosystems targeting neural network models and identifies existing solutions, analyzes potential weaknesses, and makes recommendations for further research.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46154452","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-16DOI: 10.3991/ijoe.v19i11.39277
Julián R. Camargo L., Oscar D. Flórez C., Andrés L. Jutinico
This article presents a proposal of project-based learning (PBL) as a didactic tool to meet the learning results (LR) proposed in the syllabus of the course Digital Design with Microcontrollers taught in the Electronic Engineering course of the Faculty of Engineering of the Universidad Distrital Francisco José de Caldas in Bogotá, Colombia. Students are provided with all the information related to the project, the design methodology, what is expected from the project, and how to evaluate the results of the work done. The proposed methodology shows the students that, from a practical project usually applied to the real environment, the theoretical information shown in the classroom is immediately applicable, increasing their motivation and willingness to work. In addition, another skill, teamwork, is reinforced by applying this type of teaching, since each member of the working group (three students per group) has a role in the project’s development. The low-cost development kit CY8CKIT-059 for PSoC5LP, manufactured by Infineon Technologies AG, is used in the course to apply the proposal. As a case study to demonstrate the methodology, the design of a data logger that stores humidity and temperature from a digital sensor, developed as one of several projects presented by students in recent semesters, is presented. When comparing the quantitative results obtained from the course in previous semesters with those obtained after applying the project-based learning methodology, a significant improvement can be seen: the percentage of students passing is significantly higher.
{"title":"Project-Based Learning as a Tool to Meet Learning Results: A Case Study of Teaching Microcontrollers","authors":"Julián R. Camargo L., Oscar D. Flórez C., Andrés L. Jutinico","doi":"10.3991/ijoe.v19i11.39277","DOIUrl":"https://doi.org/10.3991/ijoe.v19i11.39277","url":null,"abstract":"This article presents a proposal of project-based learning (PBL) as a didactic tool to meet the learning results (LR) proposed in the syllabus of the course Digital Design with Microcontrollers taught in the Electronic Engineering course of the Faculty of Engineering of the Universidad Distrital Francisco José de Caldas in Bogotá, Colombia. Students are provided with all the information related to the project, the design methodology, what is expected from the project, and how to evaluate the results of the work done. The proposed methodology shows the students that, from a practical project usually applied to the real environment, the theoretical information shown in the classroom is immediately applicable, increasing their motivation and willingness to work. In addition, another skill, teamwork, is reinforced by applying this type of teaching, since each member of the working group (three students per group) has a role in the project’s development. The low-cost development kit CY8CKIT-059 for PSoC5LP, manufactured by Infineon Technologies AG, is used in the course to apply the proposal. As a case study to demonstrate the methodology, the design of a data logger that stores humidity and temperature from a digital sensor, developed as one of several projects presented by students in recent semesters, is presented. When comparing the quantitative results obtained from the course in previous semesters with those obtained after applying the project-based learning methodology, a significant improvement can be seen: the percentage of students passing is significantly higher.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45001610","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.3991/ijoe.v19i10.39061
U. H. Jaid, A. Abdulhassan
The automatic estimation of speaker characteristics, such as height, age, and gender, has various applications in forensics, surveillance, customer service, and many human-robot interaction applications. These applications are often required to produce a response promptly. This work proposes a novel approach to speaker profiling by combining filter bank initializations, such as continuous wavelets and gammatone filter banks, with one-dimensional (1D) convolutional neural networks (CNN) and residual blocks. The proposed end-to-end model goes from the raw waveform to an estimated height, age, and gender of the speaker by learning speaker representation directly from the audio signal without relying on handcrafted and pre-computed acoustic features. The conducted experiments on the TIMIT dataset show that the proposed approach outperforms many previous studies on speaker profiling with a mean absolute error (MAE) of 5.18 and 4.91 cm in height estimation and MAE of 5.36 and 6.07 years in age estimation for males and females, respectively, and achieving an accuracy of 99.98% in gender prediction.
{"title":"End-to-End Speaker Profiling Using 1D CNN Architectures and Filter Bank Initialization","authors":"U. H. Jaid, A. Abdulhassan","doi":"10.3991/ijoe.v19i10.39061","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.39061","url":null,"abstract":"The automatic estimation of speaker characteristics, such as height, age, and gender, has various applications in forensics, surveillance, customer service, and many human-robot interaction applications. These applications are often required to produce a response promptly. This work proposes a novel approach to speaker profiling by combining filter bank initializations, such as continuous wavelets and gammatone filter banks, with one-dimensional (1D) convolutional neural networks (CNN) and residual blocks. The proposed end-to-end model goes from the raw waveform to an estimated height, age, and gender of the speaker by learning speaker representation directly from the audio signal without relying on handcrafted and pre-computed acoustic features. The conducted experiments on the TIMIT dataset show that the proposed approach outperforms many previous studies on speaker profiling with a mean absolute error (MAE) of 5.18 and 4.91 cm in height estimation and MAE of 5.36 and 6.07 years in age estimation for males and females, respectively, and achieving an accuracy of 99.98% in gender prediction.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48693580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Gait cycle plays a major role in human locomotion. Patients with neuromuscular problems are unable to walk normally. Foot drop causes difficulty in lifting the front part of the foot and affects the dorsiflexion (DF) and plantar flexion (PF) motion of the foot. Patient with foot drop must use ankle braces to achieve a normal gait. The existing ankle-foot orthosis (AFO) has its own limitations, as it does not produce adequate PF motion. To overcome this scenario, a study was conducted to analyse the two-degrees-of-freedom (DOF) motion of a robotic ankle foot orthosis (RAFO) with a spring-based series elastic actuator (SEA) and scissor actuator. The objective of this paper is to evaluate the two DOF of RAFO with two different actuators using simscape multibody. The RAFO with actuators were designed using Solidworks, and simulation was carried out using simscape multibody, to analyse the 2-DOF motion. The dynamic motion analysis was carried out using block libraries, bodies, joints, constraints, revolute joints, sensors and a proportional integral (PI) controller. From the simulation results, the total range of motion (ROM) 40° (PF angle of –25° and DF angle of 15°) is achieved by the proposed RAFO with different actuators. Further, based on the results, the input power consumption of spring-based SEA was found to be less than the scissor actuator. Similarly, torque and output power generation of the scissor actuator was found to be greater than spring-based SEA to achieve the normal human ROM. Hence, the designer can choose a hybrid actuator for foot-drop-disorder applications.
{"title":"Evaluation of Robotic Ankle-Foot Orthosis with Different Actuators Using Simscape Multibody for Foot-Drop Patients","authors":"Gowrishankar Govindaraj, Arockia Selvakumar Arockia Doss","doi":"10.3991/ijoe.v19i10.40375","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.40375","url":null,"abstract":"Gait cycle plays a major role in human locomotion. Patients with neuromuscular problems are unable to walk normally. Foot drop causes difficulty in lifting the front part of the foot and affects the dorsiflexion (DF) and plantar flexion (PF) motion of the foot. Patient with foot drop must use ankle braces to achieve a normal gait. The existing ankle-foot orthosis (AFO) has its own limitations, as it does not produce adequate PF motion. To overcome this scenario, a study was conducted to analyse the two-degrees-of-freedom (DOF) motion of a robotic ankle foot orthosis (RAFO) with a spring-based series elastic actuator (SEA) and scissor actuator. The objective of this paper is to evaluate the two DOF of RAFO with two different actuators using simscape multibody. The RAFO with actuators were designed using Solidworks, and simulation was carried out using simscape multibody, to analyse the 2-DOF motion. The dynamic motion analysis was carried out using block libraries, bodies, joints, constraints, revolute joints, sensors and a proportional integral (PI) controller. From the simulation results, the total range of motion (ROM) 40° (PF angle of –25° and DF angle of 15°) is achieved by the proposed RAFO with different actuators. Further, based on the results, the input power consumption of spring-based SEA was found to be less than the scissor actuator. Similarly, torque and output power generation of the scissor actuator was found to be greater than spring-based SEA to achieve the normal human ROM. Hence, the designer can choose a hybrid actuator for foot-drop-disorder applications.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49358875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.3991/ijoe.v19i10.39721
V. T. H. Tuyet, N. T. Binh
Neural networks overcome drawbacks of vision tasks by becoming convolutional in a wide range of layers. The salient map is affected by multilevels of strong pixels (superpixels) in global images and that is dependent on the hard threshold for their dividing. Deep neural networks have been established for saliency prediction of segmentation because the feature extraction must be suited to the input data. The convolutional neural network (CNN) also endures conflict between spatial pattern and a likeness of salient objects. Semantic segmentation is one of the approaches to continue classification based on these features. Therefore, upgrading the extraction process can be of use in saliency. In this work, we optimize DeepLab based on an atrous convolutional and a conditional random field (CRF) with a bottleneck in the semantic segmentation method, which serves for classification. The backbone of deep feature extraction is atrous convolution and the bottleneck based on CRF for hybrid saliency in the encoder-decoder system. The classification results are compared with some approaches for saliency prediction of recent deeper methods in an ISIC 2017 dataset. The results give better values not only for saliency prediction for segmentation but also for training and testing for classification.
{"title":"Melanoma Classification via Hybrid Saliency and Conditional Random Field with Bottleneck to Optimize DeepLab","authors":"V. T. H. Tuyet, N. T. Binh","doi":"10.3991/ijoe.v19i10.39721","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.39721","url":null,"abstract":"Neural networks overcome drawbacks of vision tasks by becoming convolutional in a wide range of layers. The salient map is affected by multilevels of strong pixels (superpixels) in global images and that is dependent on the hard threshold for their dividing. Deep neural networks have been established for saliency prediction of segmentation because the feature extraction must be suited to the input data. The convolutional neural network (CNN) also endures conflict between spatial pattern and a likeness of salient objects. Semantic segmentation is one of the approaches to continue classification based on these features. Therefore, upgrading the extraction process can be of use in saliency. In this work, we optimize DeepLab based on an atrous convolutional and a conditional random field (CRF) with a bottleneck in the semantic segmentation method, which serves for classification. The backbone of deep feature extraction is atrous convolution and the bottleneck based on CRF for hybrid saliency in the encoder-decoder system. The classification results are compared with some approaches for saliency prediction of recent deeper methods in an ISIC 2017 dataset. The results give better values not only for saliency prediction for segmentation but also for training and testing for classification.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48239119","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.3991/ijoe.v19i10.41285
Samar Raza Talpur, Huma Sikandar, Alhamzah F. Abbas, Javed Ali
A decentralised, tamper-proof ledger offered by blockchain technology has the potential to revolutionise the manufacturing sector by enhancing digital rights management, supply chain management, and product monitoring and tracking. Industrial supply chains may be made more transparent, secure, and efficient with the use of blockchain technology. This will save costs, boost quality control, and raise consumer confidence that the goods they buy are genuine and high calibre. However, there is a research gap in the implications of blockchain technology in the manufacturing sector. The aim of this research was to investigate the challenges and opportunities of blockchain technologies in the manufacturing sector. In order to accomplish the study’s goal, a two-stage systematic literature review technique was used, with the PRISMA framework being used to gather pertinent data from reliable sources like Scopus. The study contained 117 research papers, which were analysed using descriptive and scientometric methods and lysis to synthesise the literature and investigate important research clusters using the centrality and co-occurrence of keywords. The study’s conclusions point to the potential of blockchain technology to support decentralised manufacturing systems that provide risk-free and trustworthy cooperation among multiple stakeholders. The report also discusses the advantages and drawbacks of using blockchain in manufacturing and offers information on recent developments in the field of digital manufacturing that are related to blockchain technology. This study emphasises the value of blockchain technology for the industrial sector and the need for more research to fully understand its potential. Blockchain technology may help the manufacturing industry become more effective, transparent, and quality assured while also reducing costs and fostering better confidence among supply chain actors.
{"title":"Revolutionizing Manufacturing with Blockchain Technology: Opportunities and Challenges","authors":"Samar Raza Talpur, Huma Sikandar, Alhamzah F. Abbas, Javed Ali","doi":"10.3991/ijoe.v19i10.41285","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.41285","url":null,"abstract":" \u0000A decentralised, tamper-proof ledger offered by blockchain technology has the potential to revolutionise the manufacturing sector by enhancing digital rights management, supply chain management, and product monitoring and tracking. Industrial supply chains may be made more transparent, secure, and efficient with the use of blockchain technology. This will save costs, boost quality control, and raise consumer confidence that the goods they buy are genuine and high calibre. However, there is a research gap in the implications of blockchain technology in the manufacturing sector. The aim of this research was to investigate the challenges and opportunities of blockchain technologies in the manufacturing sector. In order to accomplish the study’s goal, a two-stage systematic literature review technique was used, with the PRISMA framework being used to gather pertinent data from reliable sources like Scopus. The study contained 117 research papers, which were analysed using descriptive and scientometric methods and lysis to synthesise the literature and investigate important research clusters using the centrality and co-occurrence of keywords. The study’s conclusions point to the potential of blockchain technology to support decentralised manufacturing systems that provide risk-free and trustworthy cooperation among multiple stakeholders. The report also discusses the advantages and drawbacks of using blockchain in manufacturing and offers information on recent developments in the field of digital manufacturing that are related to blockchain technology. This study emphasises the value of blockchain technology for the industrial sector and the need for more research to fully understand its potential. Blockchain technology may help the manufacturing industry become more effective, transparent, and quality assured while also reducing costs and fostering better confidence among supply chain actors.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44582202","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.3991/ijoe.v19i10.41399
Vichayanan Rattanawiboomsom, Muhammad Saleem Korejo, Javed Ali, Uthen Thatsaringkharnsakun
The use of advanced computer technology in the healthcare industry has the potential to improve patient care and therapeutic results. The goal of this project is to improve data security, privacy, and decentralisation in healthcare by integrating blockchain and Internet of Things (IoT) technologies. The adoption of IoT devices makes it possible to gather and analyse patient sensory data in real–time; however centralised processing and storage present problems such as data manipulation and privacy issues. The study investigates the creation of a decentralised IoT-based e-healthcare system that takes these issues into account by utilising blockchain technology. In addition, the paper also emphasises how blockchain use has advanced smart contract technologies. Smart contracts provide safe user authentication for IoT device access, assuring responsibility, traceability, and data integrity. The study investigates the potentially game-changing applications of blockchain technology in healthcare, such as enhanced data interoperability, patient-cantered care, reduced administrative procedures, and increased transaction transparency. The report also highlights the significance of blockchain in managing pharmaceutical supply chains, considering the essential influence on patient welfare and safety. Effective management is essential in the healthcare business because supply chain interruptions or breaches can have serious implications. The present level of research in blockchain-enabled IoT applications for healthcare is examined comprehensively using the PRISMA framework and records from the Scopus database. The three most important research topics are cloud computing, fog computing, and medical services. The results highlight the important role that blockchain-enabled IoT applications have played in enhancing data security and privacy in the healthcare industry. Real-time data gathering, precise diagnoses, individualised treatments, and simplified administrative procedures are all made possible by the integration of blockchain and IoT. Additionally, scalable solutions and insightful data for healthcare decision-making are provided via fog computing, cloud computing, machine learning, and smart contracts.
{"title":"Blockchain-Enabled Internet of Things (IoT) Applications in Healthcare: A Systematic Review of Current Trends and Future Opportunities","authors":"Vichayanan Rattanawiboomsom, Muhammad Saleem Korejo, Javed Ali, Uthen Thatsaringkharnsakun","doi":"10.3991/ijoe.v19i10.41399","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.41399","url":null,"abstract":"The use of advanced computer technology in the healthcare industry has the potential to improve patient care and therapeutic results. The goal of this project is to improve data security, privacy, and decentralisation in healthcare by integrating blockchain and Internet of Things (IoT) technologies. The adoption of IoT devices makes it possible to gather and analyse patient sensory data in real–time; however centralised processing and storage present problems such as data manipulation and privacy issues. The study investigates the creation of a decentralised IoT-based e-healthcare system that takes these issues into account by utilising blockchain technology. In addition, the paper also emphasises how blockchain use has advanced smart contract technologies. Smart contracts provide safe user authentication for IoT device access, assuring responsibility, traceability, and data integrity. The study investigates the potentially game-changing applications of blockchain technology in healthcare, such as enhanced data interoperability, patient-cantered care, reduced administrative procedures, and increased transaction transparency. The report also highlights the significance of blockchain in managing pharmaceutical supply chains, considering the essential influence on patient welfare and safety. Effective management is essential in the healthcare business because supply chain interruptions or breaches can have serious implications. The present level of research in blockchain-enabled IoT applications for healthcare is examined comprehensively using the PRISMA framework and records from the Scopus database. The three most important research topics are cloud computing, fog computing, and medical services. The results highlight the important role that blockchain-enabled IoT applications have played in enhancing data security and privacy in the healthcare industry. Real-time data gathering, precise diagnoses, individualised treatments, and simplified administrative procedures are all made possible by the integration of blockchain and IoT. Additionally, scalable solutions and insightful data for healthcare decision-making are provided via fog computing, cloud computing, machine learning, and smart contracts.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43061479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.3991/ijoe.v19i10.39583
Muhanad Abdul Elah Alkhalisy, Saad Hameed Abid
Massive open online courses (MOOCs) and other forms of distance learning have gained popularity in recent years. The success of remote online exam proctoring determines the integrity of the exam. Deep-learning-powered proctoring services have also grown in popularity. A large number of samples are needed for deep-learning training. The network’s generalization ability is poor due to insufficient training data or an uneven lack of variation. This study illustrates how to analyze students’ anomalous behavior by utilizing a YOLOv5 deep model trained using newly produced dataset. To overcome insufficient training data for deep-learning-related issues, this paper proposes a data-augmentation method based on semantic segmentation. The MobileNetV3 model was used to get an image semantic segmentation mask, which was used to get a binary mask, which in turn was used to replace the image background by using conditional subtraction with randomly selected background images. Finally, randomly pixel-based color augmentation was added to the resulting image. The behavioral detection model used in this study achieved 0.98 mean average precision (mAP) on the produced dataset, showing acceptable detection precision. The experimental findings indicate that the suggested augmentation method improves behavioral detection precision by more than 0.3%.
{"title":"Abnormal Behavior Detection in Online Exams Using Deep Learning and Data Augmentation Techniques","authors":"Muhanad Abdul Elah Alkhalisy, Saad Hameed Abid","doi":"10.3991/ijoe.v19i10.39583","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.39583","url":null,"abstract":"Massive open online courses (MOOCs) and other forms of distance learning have gained popularity in recent years. The success of remote online exam proctoring determines the integrity of the exam. Deep-learning-powered proctoring services have also grown in popularity. A large number of samples are needed for deep-learning training. The network’s generalization ability is poor due to insufficient training data or an uneven lack of variation. This study illustrates how to analyze students’ anomalous behavior by utilizing a YOLOv5 deep model trained using newly produced dataset. To overcome insufficient training data for deep-learning-related issues, this paper proposes a data-augmentation method based on semantic segmentation. The MobileNetV3 model was used to get an image semantic segmentation mask, which was used to get a binary mask, which in turn was used to replace the image background by using conditional subtraction with randomly selected background images. Finally, randomly pixel-based color augmentation was added to the resulting image. The behavioral detection model used in this study achieved 0.98 mean average precision (mAP) on the produced dataset, showing acceptable detection precision. The experimental findings indicate that the suggested augmentation method improves behavioral detection precision by more than 0.3%.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48883459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.3991/ijoe.v19i10.37681
S. Alomari
The two main causes of blindness are diabetes and glaucoma. Routine diagnosis of blindness is based on the conventional robust mass-screening method. However, despite being cost-effective, this method has some problems as a human eye-disease detection method because there are many types of eye disease that are similar or that result in no visual changes in the eye image. These issues make it highly difficult to recognize blindness and control it. Moreover, the color of the macula of the spot can be very close to that of the affected macula in a variety of eye diseases, which suggests that the color of the macula spot can indicate various possibilities, rather than one. This paper discusses the shortcomings of current blindness-screening and monitoring systems and presents a feature-based blindness diagnosis approach using digital eye fundus images for the purpose of automated diagnosis of eye disorders, considering three conditions: healthy eye, diabetic retinopathy (DR), and glaucoma. As such, this paper develops a computer-aided diagnosis (CAD) method for automated detection of human blindness. The proposed approach integrates Gabor filter features, statistical features, colored features, morphological features, and local binary pattern features, then compares them with features drawn from a standard dataset of 1580 fundus images. Several classification techniques were applied to the extracted-features neural network (NN), support vector machine (SVM), naïve bias (NB). SVM classifiers show the most promising accuracy. They achieved 93.3% over the other classifiers.
{"title":"An Efficient System for Diagnosis of Human Blindness Using Image-Processing and Machine-Learning Methods","authors":"S. Alomari","doi":"10.3991/ijoe.v19i10.37681","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.37681","url":null,"abstract":"The two main causes of blindness are diabetes and glaucoma. Routine diagnosis of blindness is based on the conventional robust mass-screening method. However, despite being cost-effective, this method has some problems as a human eye-disease detection method because there are many types of eye disease that are similar or that result in no visual changes in the eye image. These issues make it highly difficult to recognize blindness and control it. Moreover, the color of the macula of the spot can be very close to that of the affected macula in a variety of eye diseases, which suggests that the color of the macula spot can indicate various possibilities, rather than one. This paper discusses the shortcomings of current blindness-screening and monitoring systems and presents a feature-based blindness diagnosis approach using digital eye fundus images for the purpose of automated diagnosis of eye disorders, considering three conditions: healthy eye, diabetic retinopathy (DR), and glaucoma. As such, this paper develops a computer-aided diagnosis (CAD) method for automated detection of human blindness. The proposed approach integrates Gabor filter features, statistical features, colored features, morphological features, and local binary pattern features, then compares them with features drawn from a standard dataset of 1580 fundus images. Several classification techniques were applied to the extracted-features neural network (NN), support vector machine (SVM), naïve bias (NB). SVM classifiers show the most promising accuracy. They achieved 93.3% over the other classifiers.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45803875","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.3991/ijoe.v19i10.38843
H. A. Ali, Liwa Abdullah Ali, A. Vasilățeanu, N. Goga, R. Popa
BCI is a rapidly growing field within biomedical engineering as it enables a direct connection between the central nervous system and an external device. BCI detects brain signals using biosensors (electrodes) installed on the head’s scalp or implanted inside the brain. EEG is a non-invasive method for detecting and monitoring the brain’s activity. Using EEG-based BCI in the medical field can significantly help disabled people to perform daily activities. In this context, it is very important to support and enable paralysed people to interact with multimedia devices like televisions by developing suitable solutions. This paper proposes an EEG mind-controlled TV remote control system prototype. The proposed prototype uses affordable hardware components to perform its task. A quantitative questionnaire has been conducted to identify the system’s functional and non-functional requirements. The system can send four different signals to power on/off, change the channel, raise and reduce the volume of the TV. The system has been tested by 20 subjects. The testing results show that the accuracy of the system is 74.9%. Despite the system being able to control only four TV functions, the system is scalable, and more commands can be added in the future. Also, using Raspberry Pi in the system gives a great possibility to eliminate the computer and to use Raspberry Pi directly with the headset. This paper demonstrates the approach’s feasibility and opens the route for enhancing the system and using EEG-based BCI with more and different devices.
{"title":"Towards Design and Implementation of an EEG-based BCI TV Remote Control","authors":"H. A. Ali, Liwa Abdullah Ali, A. Vasilățeanu, N. Goga, R. Popa","doi":"10.3991/ijoe.v19i10.38843","DOIUrl":"https://doi.org/10.3991/ijoe.v19i10.38843","url":null,"abstract":"BCI is a rapidly growing field within biomedical engineering as it enables a direct connection between the central nervous system and an external device. BCI detects brain signals using biosensors (electrodes) installed on the head’s scalp or implanted inside the brain. EEG is a non-invasive method for detecting and monitoring the brain’s activity. Using EEG-based BCI in the medical field can significantly help disabled people to perform daily activities. In this context, it is very important to support and enable paralysed people to interact with multimedia devices like televisions by developing suitable solutions. This paper proposes an EEG mind-controlled TV remote control system prototype. The proposed prototype uses affordable hardware components to perform its task. A quantitative questionnaire has been conducted to identify the system’s functional and non-functional requirements. The system can send four different signals to power on/off, change the channel, raise and reduce the volume of the TV. The system has been tested by 20 subjects. The testing results show that the accuracy of the system is 74.9%. Despite the system being able to control only four TV functions, the system is scalable, and more commands can be added in the future. Also, using Raspberry Pi in the system gives a great possibility to eliminate the computer and to use Raspberry Pi directly with the headset. This paper demonstrates the approach’s feasibility and opens the route for enhancing the system and using EEG-based BCI with more and different devices.","PeriodicalId":36900,"journal":{"name":"International Journal of Online and Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":1.3,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44331426","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：