Pub Date : 2023-01-09DOI: 10.1109/DeSE58274.2023.10100119
M. Mahyoub, F. Natalia, S. Sudirman, A. Al-Jumaily, P. Liatsis
Early and accurate detection of brain tumors is very important to save the patient's life. Brain tumors are generally diagnosed manually by a radiologist by analyzing the patient”s brain MRI scans which is a time-consuming process. This led to our study of this research area for finding out a solution to automate the diagnosis to increase its speed and accuracy. In this study, we investigate the use of Residual Network deep learning architecture to diagnose and segment brain tumors. We proposed a two-step method involving a tumor detection stage, using ResNet50 architecture, and a tumor area segmentation stage using ResU-Net architecture. We adopt transfer learning on pre-trained models to help get the best performance out of the approach, as well as data augmentation to lessen the effect of data population imbalance and hyperparameter optimization to get the best set of training parameter values. Using a publicly available dataset as a testbed we show that our approach achieves 84.3 % performance outperforming the state-of-the-art using U-Net by 2% using the Dice Coefficient metric.
{"title":"Brain Tumor Segmentation in Fluid-Attenuated Inversion Recovery Brain MRI using Residual Network Deep Learning Architectures","authors":"M. Mahyoub, F. Natalia, S. Sudirman, A. Al-Jumaily, P. Liatsis","doi":"10.1109/DeSE58274.2023.10100119","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10100119","url":null,"abstract":"Early and accurate detection of brain tumors is very important to save the patient's life. Brain tumors are generally diagnosed manually by a radiologist by analyzing the patient”s brain MRI scans which is a time-consuming process. This led to our study of this research area for finding out a solution to automate the diagnosis to increase its speed and accuracy. In this study, we investigate the use of Residual Network deep learning architecture to diagnose and segment brain tumors. We proposed a two-step method involving a tumor detection stage, using ResNet50 architecture, and a tumor area segmentation stage using ResU-Net architecture. We adopt transfer learning on pre-trained models to help get the best performance out of the approach, as well as data augmentation to lessen the effect of data population imbalance and hyperparameter optimization to get the best set of training parameter values. Using a publicly available dataset as a testbed we show that our approach achieves 84.3 % performance outperforming the state-of-the-art using U-Net by 2% using the Dice Coefficient metric.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129467989","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10099678
Surajit Das, Subhodeep Mukherjee
In this paper, an advanced optimization technique will be used to find the cut-off of base model(s) and meta model along with the weights of the weighted blending. In this work, XGBoost, Random Forest, Logistic Regression have been used as the base model and also K-Fold cross validation has been used to capture the average score of individual base model. Here F-score will be used to assess the goodness of the models. The techniques have been applied for classification of Breast Carcinoma which is the one of the most prevailing diseases that thrives amid the human beings over decades. According to a report, published in March '21, in the web site of WHO, in 2020, about 2.3 million women diagnosed with breast cancer and according to International Agency for Research on Cancer (IARC) in December 2020, breast cancer has overtaken the lung cancer and has reached at the top position as a commonly diagnosed cancer. In order to determine the breast carcinoma, breast tumors are classified into two categories which are tagged as malignant or benign. For this study the WBCD dataset has been used as the dataset that contains 569 records derived from Fine Needle Aspirates (FNA) of human breast masses has no missing value and is a balanced dataset which minimizes the data pre-processing and EDA steps. In the Optimized weighted Blending, the F-1 Score goes maximum 0.99 (approx.) compared to other approaches within our scope.
{"title":"Advanced Optimization Techniques & Its Application in AI-Powered Breast Cancer Classification","authors":"Surajit Das, Subhodeep Mukherjee","doi":"10.1109/DeSE58274.2023.10099678","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10099678","url":null,"abstract":"In this paper, an advanced optimization technique will be used to find the cut-off of base model(s) and meta model along with the weights of the weighted blending. In this work, XGBoost, Random Forest, Logistic Regression have been used as the base model and also K-Fold cross validation has been used to capture the average score of individual base model. Here F-score will be used to assess the goodness of the models. The techniques have been applied for classification of Breast Carcinoma which is the one of the most prevailing diseases that thrives amid the human beings over decades. According to a report, published in March '21, in the web site of WHO, in 2020, about 2.3 million women diagnosed with breast cancer and according to International Agency for Research on Cancer (IARC) in December 2020, breast cancer has overtaken the lung cancer and has reached at the top position as a commonly diagnosed cancer. In order to determine the breast carcinoma, breast tumors are classified into two categories which are tagged as malignant or benign. For this study the WBCD dataset has been used as the dataset that contains 569 records derived from Fine Needle Aspirates (FNA) of human breast masses has no missing value and is a balanced dataset which minimizes the data pre-processing and EDA steps. In the Optimized weighted Blending, the F-1 Score goes maximum 0.99 (approx.) compared to other approaches within our scope.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130055727","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10099530
Krishan Sivasankaran Pillay, Kamalanathan Shanmugam, Muhammad Ehsan Rana
Flood disaster is known to impact people and the environment substantially. People impacted by floods may lose properties and homes. Furthermore, there are also a substantial number of deaths yearly in case of a flood disaster. Malaysia has been hit by floods pretty frequently for the past few years. Government bodies and non-government organisations have collaborated to develop necessary measures to mitigate the flood disaster. Flood forecasting, warning, and zoning have been acknowledged as some of the few non-structural strategies considered crucial in flood mitigation. All these measures are vital to reduce the impact of the flood disaster on the people, eventually making them prepared to embrace any emergency. Technology is believed to play an essential role in tackling the flood issue. Several countries all across the globe have relied on the current advancements in technology to make predictions regarding flood incidents, weather forecasting and so on. Moreover, these systems give warnings and alerts to the people before a flood disaster. Therefore, utilising technology in tackling disasters like floods can reduce the severe impacts of floods in the years to come and also tends to end the loss of lives due to floods. As part of this research, the authors have first investigated the requirements and then proposed the design of a simple and easy-to-implement IoT-based flood monitoring system. Finally, a prototype is prepared to provide the proof of concept of the proposed solution.
{"title":"Recommendations for Developing an Affordable IoT-Based Flood Monitoring and Early Warning System","authors":"Krishan Sivasankaran Pillay, Kamalanathan Shanmugam, Muhammad Ehsan Rana","doi":"10.1109/DeSE58274.2023.10099530","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10099530","url":null,"abstract":"Flood disaster is known to impact people and the environment substantially. People impacted by floods may lose properties and homes. Furthermore, there are also a substantial number of deaths yearly in case of a flood disaster. Malaysia has been hit by floods pretty frequently for the past few years. Government bodies and non-government organisations have collaborated to develop necessary measures to mitigate the flood disaster. Flood forecasting, warning, and zoning have been acknowledged as some of the few non-structural strategies considered crucial in flood mitigation. All these measures are vital to reduce the impact of the flood disaster on the people, eventually making them prepared to embrace any emergency. Technology is believed to play an essential role in tackling the flood issue. Several countries all across the globe have relied on the current advancements in technology to make predictions regarding flood incidents, weather forecasting and so on. Moreover, these systems give warnings and alerts to the people before a flood disaster. Therefore, utilising technology in tackling disasters like floods can reduce the severe impacts of floods in the years to come and also tends to end the loss of lives due to floods. As part of this research, the authors have first investigated the requirements and then proposed the design of a simple and easy-to-implement IoT-based flood monitoring system. Finally, a prototype is prepared to provide the proof of concept of the proposed solution.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128922033","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10099504
M. Mahyoub, F. Natalia, S. Sudirman, A. Al-Jumaily, P. Liatsis
In autonomous driving, environment perception is an important step in understanding the driving scene. Objects in images captured through a vehicle camera can be detected and classified using semantic segmentation and depth estimation methods. Both these tasks are closely related to each other and this association helps in building a multi-task neural network where a single network is used to generate both views from a given monocular image. This approach gives the flexibility to include multiple related tasks in a single network. It helps reduce multiple independent networks and improve the performance of all related tasks. The main aim of our research presented in this paper is to build a multi-task deep learning network for simultaneous semantic segmentation and depth estimation from monocular images. Two decoder-focused U-N et-based multi-task networks that use a pre-trained Resnet-50 and DenseNet-121 which shared encoder and task-specific decoder networks with Attention Mechanisms are considered. We also employed multi-task optimization strategies such as equal weighting and dynamic weight averaging during the training of the models. The corresponding models' performance is evaluated using mean IoU for semantic segmentation and Root Mean Square Error for depth estimation. From our experiments, we found that the performance of these multi-task networks is on par with the corresponding single-task networks.
{"title":"Semantic Segmentation and Depth Estimation of Urban Road Scene Images Using Multi-Task Networks","authors":"M. Mahyoub, F. Natalia, S. Sudirman, A. Al-Jumaily, P. Liatsis","doi":"10.1109/DeSE58274.2023.10099504","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10099504","url":null,"abstract":"In autonomous driving, environment perception is an important step in understanding the driving scene. Objects in images captured through a vehicle camera can be detected and classified using semantic segmentation and depth estimation methods. Both these tasks are closely related to each other and this association helps in building a multi-task neural network where a single network is used to generate both views from a given monocular image. This approach gives the flexibility to include multiple related tasks in a single network. It helps reduce multiple independent networks and improve the performance of all related tasks. The main aim of our research presented in this paper is to build a multi-task deep learning network for simultaneous semantic segmentation and depth estimation from monocular images. Two decoder-focused U-N et-based multi-task networks that use a pre-trained Resnet-50 and DenseNet-121 which shared encoder and task-specific decoder networks with Attention Mechanisms are considered. We also employed multi-task optimization strategies such as equal weighting and dynamic weight averaging during the training of the models. The corresponding models' performance is evaluated using mean IoU for semantic segmentation and Root Mean Square Error for depth estimation. From our experiments, we found that the performance of these multi-task networks is on par with the corresponding single-task networks.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126914639","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10099871
M. Mahyoub, S. Abdulhussain, F. Natalia, S. Sudirman, Basheera M. Mahmmod
Abstract pattern is very commonly used in the textile and fashion industry. Pattern design is an area where designers need to come up with new and attractive patterns every day. It is very difficult to find employees with a sufficient creative mindset and the necessary skills to come up with new unseen attractive designs. Therefore, it would be ideal to identify a process that would allow for these patterns to be generated on their own with little to no human interaction. This can be achieved using deep learning models and techniques. One of the most recent and promising tools to solve this type of problem is Generative Adversarial Networks (GANs). In this paper, we investigate the suitability of GAN in producing abstract patterns. We achieve this by generating abstract design patterns using the two most popular GANs, namely Deep Convolutional GAN and Wasserstein GAN. By identifying the best-performing model after training using hyperparameter optimization and generating some output patterns we show that Wasserstein GAN is superior to Deep Convolutional GAN.
{"title":"Abstract Pattern Image Generation using Generative Adversarial Networks","authors":"M. Mahyoub, S. Abdulhussain, F. Natalia, S. Sudirman, Basheera M. Mahmmod","doi":"10.1109/DeSE58274.2023.10099871","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10099871","url":null,"abstract":"Abstract pattern is very commonly used in the textile and fashion industry. Pattern design is an area where designers need to come up with new and attractive patterns every day. It is very difficult to find employees with a sufficient creative mindset and the necessary skills to come up with new unseen attractive designs. Therefore, it would be ideal to identify a process that would allow for these patterns to be generated on their own with little to no human interaction. This can be achieved using deep learning models and techniques. One of the most recent and promising tools to solve this type of problem is Generative Adversarial Networks (GANs). In this paper, we investigate the suitability of GAN in producing abstract patterns. We achieve this by generating abstract design patterns using the two most popular GANs, namely Deep Convolutional GAN and Wasserstein GAN. By identifying the best-performing model after training using hyperparameter optimization and generating some output patterns we show that Wasserstein GAN is superior to Deep Convolutional GAN.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127207761","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10099815
Brentton Wong Swee Kit, M. Joseph
Voice of the Customer (VoC) has gained traction over the past few years to understand the consumers' opinion, preferences, and expectation. Reviews that are posted online are one of the methods of communication between the company and the consumers. Therefore, companies can analyse the reviews posted online to identify the aspects and sentiments that are mentioned in the reviews. However, the process of analysing the reviews manually is inefficient and is prone to bias. One of the methods of tackling manually analysing is by using machine learning. This process is called aspect-based sentiment analysis, there are many aspect-based sentiment analysis studies and research has been done previously. However, majority of the previous studies focuses on other domains such as product reviews or restaurant reviews. Therefore, this research will focus on the movie industry where movie reviews will be used to train and predict the aspects and sentiment of the movie review using machine learning models. This research will perform both aspect prediction and sentiment prediction on different models. The aspect prediction will be done using Logistic Regression and Decision Tree whist the Sentiment Analysis will be done using Logistic Regression and Multinomial Naïve Bayes. Based on the findings of the study, Decision Tree was able to achieve a higher accuracy of 98% while Logistic Regression was able to score an accuracy of 92%. Additionally, Logistic Regression was able to score a better accuracy for Sentiment Prediction with an accuracy of 93% when compared to Multinomial Naïve Bayes which achieved an accuracy of 91 %. Therefore, Decision Tree is more suitable for Aspect Prediction whilst Logistic Regression is more suitable for Sentiment Analysis.
{"title":"Aspect-Based Sentiment Analysis on Movie Reviews","authors":"Brentton Wong Swee Kit, M. Joseph","doi":"10.1109/DeSE58274.2023.10099815","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10099815","url":null,"abstract":"Voice of the Customer (VoC) has gained traction over the past few years to understand the consumers' opinion, preferences, and expectation. Reviews that are posted online are one of the methods of communication between the company and the consumers. Therefore, companies can analyse the reviews posted online to identify the aspects and sentiments that are mentioned in the reviews. However, the process of analysing the reviews manually is inefficient and is prone to bias. One of the methods of tackling manually analysing is by using machine learning. This process is called aspect-based sentiment analysis, there are many aspect-based sentiment analysis studies and research has been done previously. However, majority of the previous studies focuses on other domains such as product reviews or restaurant reviews. Therefore, this research will focus on the movie industry where movie reviews will be used to train and predict the aspects and sentiment of the movie review using machine learning models. This research will perform both aspect prediction and sentiment prediction on different models. The aspect prediction will be done using Logistic Regression and Decision Tree whist the Sentiment Analysis will be done using Logistic Regression and Multinomial Naïve Bayes. Based on the findings of the study, Decision Tree was able to achieve a higher accuracy of 98% while Logistic Regression was able to score an accuracy of 92%. Additionally, Logistic Regression was able to score a better accuracy for Sentiment Prediction with an accuracy of 93% when compared to Multinomial Naïve Bayes which achieved an accuracy of 91 %. Therefore, Decision Tree is more suitable for Aspect Prediction whilst Logistic Regression is more suitable for Sentiment Analysis.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132380396","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10100235
Ghufran Jafar Salman, M. S. M. Altaei
Building a system to recognize Arabic words or texts has been challenging. It's harder when the text is in various sizes and fonts, regardless of font complexities. This work built a smart system to recognize Arabic words and texts by creating a dataset and training it by using deep learning techniques. This system can scan text into a computer texts. Each of the 1,000 words in the dataset was written out 24 different ways, using 24 different Arabic fonts. Words in images were identified and deduced with the use of image processing methods. Finally, the deep learning (Convolution Neural Network CNN) algorithm takes over, extracting features from the truncated word and retrieving text words that are visually similar to the ones that were cut. In experiments, the system achieved 99% accuracy in words detection and 96% accuracy in recognition.
{"title":"Proposed Deep Learning System for Arabic Text Detection and Recognition","authors":"Ghufran Jafar Salman, M. S. M. Altaei","doi":"10.1109/DeSE58274.2023.10100235","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10100235","url":null,"abstract":"Building a system to recognize Arabic words or texts has been challenging. It's harder when the text is in various sizes and fonts, regardless of font complexities. This work built a smart system to recognize Arabic words and texts by creating a dataset and training it by using deep learning techniques. This system can scan text into a computer texts. Each of the 1,000 words in the dataset was written out 24 different ways, using 24 different Arabic fonts. Words in images were identified and deduced with the use of image processing methods. Finally, the deep learning (Convolution Neural Network CNN) algorithm takes over, extracting features from the truncated word and retrieving text words that are visually similar to the ones that were cut. In experiments, the system achieved 99% accuracy in words detection and 96% accuracy in recognition.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"43 21","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132736918","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10100050
Shyamala Rajasekar
Predictive maintenance is one of the main trends noted in Industry 4.0, the ongoing era of automation and digitization in the manufacturing sector. Condition monitoring, a widely prevalent technique in Predictive Maintenance involves constant monitoring of systems through sensors and technologies enabling timely intervention to prevent sudden/unplanned breakdown that affects production, man-hours, inventory, and in worst cases, safety. This paper uses a data-driven approach to identify and classify faults in a multi-component hydraulic rig. Different Feature extraction/selection methods from the historical data with multiple sensor readings of different sampling frequencies (asynchronous data) were explored and compared. Supervised Learning models were built on these features to distinguish and detect the different levels of components' degradation. In addition, given the challenge of lack of annotated data in Industrial setups, unsupervised Clustering and anomaly detection algorithms were also examined to detect faults in the system.
{"title":"Prediction of Component Level Degradation in a Hydraulic Rig using Machine Learning Methods","authors":"Shyamala Rajasekar","doi":"10.1109/DeSE58274.2023.10100050","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10100050","url":null,"abstract":"Predictive maintenance is one of the main trends noted in Industry 4.0, the ongoing era of automation and digitization in the manufacturing sector. Condition monitoring, a widely prevalent technique in Predictive Maintenance involves constant monitoring of systems through sensors and technologies enabling timely intervention to prevent sudden/unplanned breakdown that affects production, man-hours, inventory, and in worst cases, safety. This paper uses a data-driven approach to identify and classify faults in a multi-component hydraulic rig. Different Feature extraction/selection methods from the historical data with multiple sensor readings of different sampling frequencies (asynchronous data) were explored and compared. Supervised Learning models were built on these features to distinguish and detect the different levels of components' degradation. In addition, given the challenge of lack of annotated data in Industrial setups, unsupervised Clustering and anomaly detection algorithms were also examined to detect faults in the system.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132208386","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10099999
N. E. AL-Qaisy, Bilal R. Al-Kaseem, Yousif Al-Dunainawi
Recently, there has been a remarkable interest in sign language recognition techniques. Especially in the field of sensor-based besides the extensive employment of open-source platforms in research and development testbeds. Sign language recognition has attracted considerable attention from academic scholars and the industry because deafness recognized as a severe and worldwide health concern. However, most studies in recognition have only focused on vision-based or image-based systems that were not suitable for outdoor usage and lack mobility features. This paper introduces a smart glove that is based on wearable sensors to achieve portable standalone system working in a real-time environment with a user-friendly interface. The presented system utilized modern approaches to collect and generate new datasets using two kinds of sensors only. This dataset was employed to develop an artificial neural network (ANN) model that was capable of predicting the alphabetic letters based on hand gestures and orientation. The ANN model was trained using Scaled Conjugate Gradient (SCG) algorithm. The obtained results showed a remarkable performance in terms of ANN accuracy for both Arabic Sign Language (ArSL) and American Sign Language (ASL) which were 96%, 98% respectively. The performance of the developed ANN model ensured its usability in real-time scenario.
{"title":"AI-Based Portable Gesture Recognition System for Hearing Impaired People Using Wearable Sensors","authors":"N. E. AL-Qaisy, Bilal R. Al-Kaseem, Yousif Al-Dunainawi","doi":"10.1109/DeSE58274.2023.10099999","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10099999","url":null,"abstract":"Recently, there has been a remarkable interest in sign language recognition techniques. Especially in the field of sensor-based besides the extensive employment of open-source platforms in research and development testbeds. Sign language recognition has attracted considerable attention from academic scholars and the industry because deafness recognized as a severe and worldwide health concern. However, most studies in recognition have only focused on vision-based or image-based systems that were not suitable for outdoor usage and lack mobility features. This paper introduces a smart glove that is based on wearable sensors to achieve portable standalone system working in a real-time environment with a user-friendly interface. The presented system utilized modern approaches to collect and generate new datasets using two kinds of sensors only. This dataset was employed to develop an artificial neural network (ANN) model that was capable of predicting the alphabetic letters based on hand gestures and orientation. The ANN model was trained using Scaled Conjugate Gradient (SCG) algorithm. The obtained results showed a remarkable performance in terms of ANN accuracy for both Arabic Sign Language (ArSL) and American Sign Language (ASL) which were 96%, 98% respectively. The performance of the developed ANN model ensured its usability in real-time scenario.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127681555","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 : 2023-01-09DOI: 10.1109/DeSE58274.2023.10099697
Shatha Ghareeb, M. Mahyoub, J. Mustafina
Phishing website detection is the task of classifying websites as phishing or legitimate based on URL parameters and certain behaviour of the site. In today's world, dependency on websites has become inevitable. With the increase in website users population and the rise of the internet, cyber-attacks have become a common thing. Attackers across the globe target innocent users to steal their personal classified information such as login credentials, credit or debit card information, which may lead to serious monetary and identity damage for the users. One of the main challenges with this problem is the constant change in phishing URLs. Due to this, there is a constant need to update the detection mechanism, which may be extinct in a short period of time. Most of the current phishing detection tools utilise the black box method, where phishing URLs are stored and queried for verification. This may not be an efficient way due to the constant change in the URLs. In this study, a machine learning based approach is proposed along with a feature selection method to select the right set of features that may contribute to higher detection accuracy. The proposed model is also aimed at being simple, faster, and interpretable. Efficiency, accuracy, and model execution time will be evaluated against the final model.
{"title":"Analysis of Feature Selection and Phishing Website Classification Using Machine Learning","authors":"Shatha Ghareeb, M. Mahyoub, J. Mustafina","doi":"10.1109/DeSE58274.2023.10099697","DOIUrl":"https://doi.org/10.1109/DeSE58274.2023.10099697","url":null,"abstract":"Phishing website detection is the task of classifying websites as phishing or legitimate based on URL parameters and certain behaviour of the site. In today's world, dependency on websites has become inevitable. With the increase in website users population and the rise of the internet, cyber-attacks have become a common thing. Attackers across the globe target innocent users to steal their personal classified information such as login credentials, credit or debit card information, which may lead to serious monetary and identity damage for the users. One of the main challenges with this problem is the constant change in phishing URLs. Due to this, there is a constant need to update the detection mechanism, which may be extinct in a short period of time. Most of the current phishing detection tools utilise the black box method, where phishing URLs are stored and queried for verification. This may not be an efficient way due to the constant change in the URLs. In this study, a machine learning based approach is proposed along with a feature selection method to select the right set of features that may contribute to higher detection accuracy. The proposed model is also aimed at being simple, faster, and interpretable. Efficiency, accuracy, and model execution time will be evaluated against the final model.","PeriodicalId":346847,"journal":{"name":"2023 15th International Conference on Developments in eSystems Engineering (DeSE)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125877234","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：