Pub Date : 2021-12-01DOI: 10.1109/uemcon53757.2021.9666562
Koushik Sutradhar, Sourav Sutradhar, Iqbal Ahmed Jhimel, S. Gupta, Mohammad Monirujjaman Khan
Stock market price prediction is a difficult undertaking that generally requires a lot of human-computer interaction. The stock market process is fraught with risk and is influenced by a variety of factors. Of all the market sectors, it is one of the most volatile and active. When buying and selling stocks from various corporations and businesses, more caution is required. As a result, stock market forecasting is an important endeavor in business and finance. This study analyzes one of the explicit forecasting tactics based on Machine Learning architectures and predictive algorithms and gives an independent model-based strategy for predicting stock prices. The predictor model is based on the Recurrent Neural Networks' LSTM (Long Short-Term Memory) architecture, which specializes in time series data classification and prediction. This model does rigorous mathematical analysis and estimates RMSE to improve forecast accuracy (Root Mean Square Error).All calculations and performance checks are done in Python 3. A number of machine learning libraries are used for prediction and visualization. This study demonstrates that stock performance, sentiment, and social data are all closely related to recent historical data, and it establishes a framework and predicts trading pattern linkages that are suited for High Frequency Stock Trading based on preset parameters using Machine Learning.
{"title":"Stock Market Prediction using Recurrent Neural Network’s LSTM Architecture","authors":"Koushik Sutradhar, Sourav Sutradhar, Iqbal Ahmed Jhimel, S. Gupta, Mohammad Monirujjaman Khan","doi":"10.1109/uemcon53757.2021.9666562","DOIUrl":"https://doi.org/10.1109/uemcon53757.2021.9666562","url":null,"abstract":"Stock market price prediction is a difficult undertaking that generally requires a lot of human-computer interaction. The stock market process is fraught with risk and is influenced by a variety of factors. Of all the market sectors, it is one of the most volatile and active. When buying and selling stocks from various corporations and businesses, more caution is required. As a result, stock market forecasting is an important endeavor in business and finance. This study analyzes one of the explicit forecasting tactics based on Machine Learning architectures and predictive algorithms and gives an independent model-based strategy for predicting stock prices. The predictor model is based on the Recurrent Neural Networks' LSTM (Long Short-Term Memory) architecture, which specializes in time series data classification and prediction. This model does rigorous mathematical analysis and estimates RMSE to improve forecast accuracy (Root Mean Square Error).All calculations and performance checks are done in Python 3. A number of machine learning libraries are used for prediction and visualization. This study demonstrates that stock performance, sentiment, and social data are all closely related to recent historical data, and it establishes a framework and predicts trading pattern linkages that are suited for High Frequency Stock Trading based on preset parameters using Machine Learning.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115478946","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 : 2021-12-01DOI: 10.1109/uemcon53757.2021.9666601
Shrikant S. Jadhav, C. Gloster, Jannatun Naher, C. Doss, Youngsoo Kim
In this paper, we present an FPGA-based multi-memory controller for accelerating computationally intensive applications. Our architecture accepts multiple inputs and produces multiple outputs for each clock cycle. The architecture includes processor cores with pipelined functional units tailored for each application. Additionally, we present an approach to achieve one to two orders-of-magnitude speedup over a traditional software implementation executing on a conventional multi-core processor. Even though the clock frequency of the Field-Programmable Custom Computing Machine (FCCM) is an order-of-magnitude slower than a conventional multi-core processor, the FCCM is significantly faster. We used the Power function as an application to demonstrate the merits of our FCCM. In our experiments, we executed the Power function in software and compared the software execution times with the execution time of an FCCM. Additionally, we also compared FCCM execution time with the OpenMP implementation of the function. Our experiments show that the results obtained using our multi-memory architecture are 57X faster than software implementation and 17X faster than OpenMP implementation executing the Power function, respectively.
{"title":"A Multi-Memory Field-Programmable Custom Computing Machine for Accelerating Compute-Intensive Applications","authors":"Shrikant S. Jadhav, C. Gloster, Jannatun Naher, C. Doss, Youngsoo Kim","doi":"10.1109/uemcon53757.2021.9666601","DOIUrl":"https://doi.org/10.1109/uemcon53757.2021.9666601","url":null,"abstract":"In this paper, we present an FPGA-based multi-memory controller for accelerating computationally intensive applications. Our architecture accepts multiple inputs and produces multiple outputs for each clock cycle. The architecture includes processor cores with pipelined functional units tailored for each application. Additionally, we present an approach to achieve one to two orders-of-magnitude speedup over a traditional software implementation executing on a conventional multi-core processor. Even though the clock frequency of the Field-Programmable Custom Computing Machine (FCCM) is an order-of-magnitude slower than a conventional multi-core processor, the FCCM is significantly faster. We used the Power function as an application to demonstrate the merits of our FCCM. In our experiments, we executed the Power function in software and compared the software execution times with the execution time of an FCCM. Additionally, we also compared FCCM execution time with the OpenMP implementation of the function. Our experiments show that the results obtained using our multi-memory architecture are 57X faster than software implementation and 17X faster than OpenMP implementation executing the Power function, respectively.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125075835","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 : 2021-12-01DOI: 10.1109/uemcon53757.2021.9666525
J. D. L. Cruz, Douglas Shimizu, K. George
Stress in various amounts has the potential to reduce the efficiency of one’s ability to perform various tasks. Heat stress specifically is a natural element that is often experienced by firefighters while they are on duty, due to both the environments they are exposed to, and the heavy protective gear that they wear. This study analyzed a subject’s stress levels using fNIRS and EEG while they played a PC game that tested their short-term memory. Trials were conducted and compared while subjects both wore and did not wear turnout firefighter gear. Heart rate, blood oxygen levels, and body temperature were also measured. EEG and fNIRS data were analyzed and processed via MATLAB. The data indicates that although stress was experienced when tested against a memory game, performance of short-term memory was not substantially impaired by it. The results of the gear and no gear trials were compared and indicated that wearing gear slightly amplified the amount of stress that was felt when testing short-term memory, although it also did not have a significantly detrimental impact on memory.
{"title":"Using EEG and fNIRS Measurements for Analysis on the Effects of Heat Stress on Short-term Memory Performance","authors":"J. D. L. Cruz, Douglas Shimizu, K. George","doi":"10.1109/uemcon53757.2021.9666525","DOIUrl":"https://doi.org/10.1109/uemcon53757.2021.9666525","url":null,"abstract":"Stress in various amounts has the potential to reduce the efficiency of one’s ability to perform various tasks. Heat stress specifically is a natural element that is often experienced by firefighters while they are on duty, due to both the environments they are exposed to, and the heavy protective gear that they wear. This study analyzed a subject’s stress levels using fNIRS and EEG while they played a PC game that tested their short-term memory. Trials were conducted and compared while subjects both wore and did not wear turnout firefighter gear. Heart rate, blood oxygen levels, and body temperature were also measured. EEG and fNIRS data were analyzed and processed via MATLAB. The data indicates that although stress was experienced when tested against a memory game, performance of short-term memory was not substantially impaired by it. The results of the gear and no gear trials were compared and indicated that wearing gear slightly amplified the amount of stress that was felt when testing short-term memory, although it also did not have a significantly detrimental impact on memory.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125864604","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 : 2021-12-01DOI: 10.1109/UEMCON53757.2021.9666547
Lama AlAwlaqi, Amaal AlDawod, Ray AlFowzan, Lamia Al-Braheem
Fog/Edge computing architectures have become hot research issues with the recent development in the Internet of Things field. Although several studies have been published in this field, there is a need to focus more on exploring the analytical techniques used with these architectures. The problem that needs to be addressed is that ignoring IoT requirements when selecting the analytical techniques may affect the performance of Fog/Edge computing. Therefore, this paper first briefly discusses the IoT requirements for Fog/Edge computing. Then, the studies related to Fog/Edge computing are presented. Moreover, a comparative analysis is conducted in order to know if the proposed architecture considers the IoT requirements or not. This can be considered as a step toward designing the efficient architecture of IoT Fog/Edge computing. In addition, highlighting the IoT requirements that are not considered may encourage researchers to contribute more to this field.
{"title":"The Requirements of Fog/Edge Computing-Based IoT Architecture","authors":"Lama AlAwlaqi, Amaal AlDawod, Ray AlFowzan, Lamia Al-Braheem","doi":"10.1109/UEMCON53757.2021.9666547","DOIUrl":"https://doi.org/10.1109/UEMCON53757.2021.9666547","url":null,"abstract":"Fog/Edge computing architectures have become hot research issues with the recent development in the Internet of Things field. Although several studies have been published in this field, there is a need to focus more on exploring the analytical techniques used with these architectures. The problem that needs to be addressed is that ignoring IoT requirements when selecting the analytical techniques may affect the performance of Fog/Edge computing. Therefore, this paper first briefly discusses the IoT requirements for Fog/Edge computing. Then, the studies related to Fog/Edge computing are presented. Moreover, a comparative analysis is conducted in order to know if the proposed architecture considers the IoT requirements or not. This can be considered as a step toward designing the efficient architecture of IoT Fog/Edge computing. In addition, highlighting the IoT requirements that are not considered may encourage researchers to contribute more to this field.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"120 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115528267","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 : 2021-12-01DOI: 10.1109/uemcon53757.2021.9666619
Saikat Das, Ph.D., S. Shiva
We have developed a Distributed Denial of Service (DDoS) intrusion detection framework that employs ML ensembles of both supervised and unsupervised classifiers that are complementary in reaching a corroborated classification decision. Our work has been limited to DDoS attack detection techniques. We propose to extend our framework to general ML system development, based on our review of current ML system development life cycles. We also propose to augment the general life cycle model to include security features to enable building security-in as the development progresses and bolt security-on as flaws are discovered after deployment. Most ML systems today operate in a black-box mode, providing users with only the predictions without associated reasoning as to how the predictions are brought about. There is heavy emphasis now to build mechanisms that help the user develop higher confidence in accepting the predictions of ML systems. Such explainability feature of ML model predictions is a must for critical systems. We also propose to augment our lifecycle model with explainability features. Thus, our ultimate goal is to develop a generic ML lifecycle process augmented with security and explainability features. Such an ML lifecycle process will be of immense use in ML systems development for all domains.
{"title":"Machine Learning application lifecycle augmented with explanation and security","authors":"Saikat Das, Ph.D., S. Shiva","doi":"10.1109/uemcon53757.2021.9666619","DOIUrl":"https://doi.org/10.1109/uemcon53757.2021.9666619","url":null,"abstract":"We have developed a Distributed Denial of Service (DDoS) intrusion detection framework that employs ML ensembles of both supervised and unsupervised classifiers that are complementary in reaching a corroborated classification decision. Our work has been limited to DDoS attack detection techniques. We propose to extend our framework to general ML system development, based on our review of current ML system development life cycles. We also propose to augment the general life cycle model to include security features to enable building security-in as the development progresses and bolt security-on as flaws are discovered after deployment. Most ML systems today operate in a black-box mode, providing users with only the predictions without associated reasoning as to how the predictions are brought about. There is heavy emphasis now to build mechanisms that help the user develop higher confidence in accepting the predictions of ML systems. Such explainability feature of ML model predictions is a must for critical systems. We also propose to augment our lifecycle model with explainability features. Thus, our ultimate goal is to develop a generic ML lifecycle process augmented with security and explainability features. Such an ML lifecycle process will be of immense use in ML systems development for all domains.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"132 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116402782","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 : 2021-12-01DOI: 10.1109/UEMCON53757.2021.9666627
Fatima Salahdine, Zakaria El Mrabet, N. Kaabouch
Phishing attacks are one of the most common social engineering attacks targeting users’ emails to fraudulently steal confidential and sensitive information. They can be used as a part of more massive attacks launched to gain a foothold in corporate or government networks. Over the last decade, a number of anti-phishing techniques have been proposed to detect and mitigate these attacks. However, they are still inefficient and inaccurate. Thus, there is a great need for efficient and accurate detection techniques to cope with these attacks. In this paper, we proposed a phishing attack detection technique based on machine learning. We collected and analyzed more than 4000 phishing emails targeting the email service of the University of North Dakota. We modeled these attacks by selecting 10 relevant features and building a large dataset. This dataset was used to train, validate, and test the machine learning algorithms. For performance evaluation, four metrics have been used, namely probability of detection, probability of miss-detection, probability of false alarm, and accuracy. The experimental results show that better detection can be achieved using an artificial neural network.
{"title":"Phishing Attacks Detection A Machine Learning-Based Approach","authors":"Fatima Salahdine, Zakaria El Mrabet, N. Kaabouch","doi":"10.1109/UEMCON53757.2021.9666627","DOIUrl":"https://doi.org/10.1109/UEMCON53757.2021.9666627","url":null,"abstract":"Phishing attacks are one of the most common social engineering attacks targeting users’ emails to fraudulently steal confidential and sensitive information. They can be used as a part of more massive attacks launched to gain a foothold in corporate or government networks. Over the last decade, a number of anti-phishing techniques have been proposed to detect and mitigate these attacks. However, they are still inefficient and inaccurate. Thus, there is a great need for efficient and accurate detection techniques to cope with these attacks. In this paper, we proposed a phishing attack detection technique based on machine learning. We collected and analyzed more than 4000 phishing emails targeting the email service of the University of North Dakota. We modeled these attacks by selecting 10 relevant features and building a large dataset. This dataset was used to train, validate, and test the machine learning algorithms. For performance evaluation, four metrics have been used, namely probability of detection, probability of miss-detection, probability of false alarm, and accuracy. The experimental results show that better detection can be achieved using an artificial neural network.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128730251","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 : 2021-12-01DOI: 10.1109/UEMCON53757.2021.9666498
Federica Aveta, Siu Man Chan, Nabil Asfari, H. Refai
Atmospheric turbulence can heavily affect free space optical communication (FSOC) link reliability. This introduces random fluctuations of the received signal intensity, resulting in degraded system communication performance. While extensive research has been conducted to estimate atmospheric turbulence on single user FSOC, the effects of turbulent channel on multi-point FSOC has recently gained attention. In fact, latest results showed the feasibility of multi-user FSOC when users, sharing time and bandwidth resources, communicate with a single optical access node. This paper presents a machine learning (ML)-based methodology to identify how many users are concurrently transmitting and overlapping into a single receiver interfering within each other, and which one is propagating through a turbulent channel. The proposed methodology presents two different approaches based on: 1) traditional classification ML algorithms and 2) Convolutional Neural Network (CNN). Both methods employ amplitude distribution of the received mixed signals as input features. 100% validation accuracy was achieved by CNN employing an experimental data set of 900 images.
{"title":"Atmospheric Turbulence Identification in a multi-user FSOC using Supervised Machine Learning","authors":"Federica Aveta, Siu Man Chan, Nabil Asfari, H. Refai","doi":"10.1109/UEMCON53757.2021.9666498","DOIUrl":"https://doi.org/10.1109/UEMCON53757.2021.9666498","url":null,"abstract":"Atmospheric turbulence can heavily affect free space optical communication (FSOC) link reliability. This introduces random fluctuations of the received signal intensity, resulting in degraded system communication performance. While extensive research has been conducted to estimate atmospheric turbulence on single user FSOC, the effects of turbulent channel on multi-point FSOC has recently gained attention. In fact, latest results showed the feasibility of multi-user FSOC when users, sharing time and bandwidth resources, communicate with a single optical access node. This paper presents a machine learning (ML)-based methodology to identify how many users are concurrently transmitting and overlapping into a single receiver interfering within each other, and which one is propagating through a turbulent channel. The proposed methodology presents two different approaches based on: 1) traditional classification ML algorithms and 2) Convolutional Neural Network (CNN). Both methods employ amplitude distribution of the received mixed signals as input features. 100% validation accuracy was achieved by CNN employing an experimental data set of 900 images.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124663411","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 : 2021-12-01DOI: 10.1109/uemcon53757.2021.9666736
Imran Al Muneem, Sakif Md. Fahim, Fazle Rabby Khan, Tanjir Ahmed Emon, Md. Sabiul Islam, Mohammad Monirujjaman Khan
Technology in the unmanned aerial vehicle (UAV) can solve many emergency problems in civilian and military sectors by doing the proper implementation [1]. However, this is not commercially used on large scale till now in many countries because of many security factors. Proper implementation of the drone can utilize the problem of emergency medical goods delivery on inaccessible roads, quick surveillance for military and government law enforcement agencies, and much more. Traditional transportation infrastructure might be affected in the same way by delivery drones. This paper is about a new drone model. With the suggested design of a drone, multipurpose work including emergency delivery and surveillance network will facilitate more time efficiency and much more economical to potentially save lives. Moreover, in the current COVID-19 pandemic situation, it will be very helpful to supply medicine and goods to the lockdown areas.
{"title":"Research and Development of Multipurpose Unmanned Aerial Vehicle (Flying Drone)","authors":"Imran Al Muneem, Sakif Md. Fahim, Fazle Rabby Khan, Tanjir Ahmed Emon, Md. Sabiul Islam, Mohammad Monirujjaman Khan","doi":"10.1109/uemcon53757.2021.9666736","DOIUrl":"https://doi.org/10.1109/uemcon53757.2021.9666736","url":null,"abstract":"Technology in the unmanned aerial vehicle (UAV) can solve many emergency problems in civilian and military sectors by doing the proper implementation [1]. However, this is not commercially used on large scale till now in many countries because of many security factors. Proper implementation of the drone can utilize the problem of emergency medical goods delivery on inaccessible roads, quick surveillance for military and government law enforcement agencies, and much more. Traditional transportation infrastructure might be affected in the same way by delivery drones. This paper is about a new drone model. With the suggested design of a drone, multipurpose work including emergency delivery and surveillance network will facilitate more time efficiency and much more economical to potentially save lives. Moreover, in the current COVID-19 pandemic situation, it will be very helpful to supply medicine and goods to the lockdown areas.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130503247","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 : 2021-12-01DOI: 10.1109/UEMCON53757.2021.9666640
Jon Jenkins, S. Rashad
Millions of people globally use some form of sign language in their everyday lives. There is a need for a method of gesture recognition that is as easy to use and ubiquitous as voice recognition is today. In this paper we explore a way to translate from sign language to speech using an innovative method, utilizing the Leap Motion Controller and machine learning algorithms to capture and analyze hand movements in real time, then converting the interpreted signs into spoken word. We seek to build a system that is easy to use, intuitive to understand, adaptable to the individual, and usable in everyday life. This system will be able to work in an adaptive way to learn new signs to expand the dictionary of the system and allow higher accuracy on an individual level. It will have a wide range of applications for healthcare, education, gamification, communication, and more. An optical hand tracking piece of hardware, the Leap Motion Controller will be used to capture hand movements and information to create supervised machine learning models that can be trained to accurately guess American Sign Language (ASL) symbols being signed in real time. Experimental results show that the proposed method is promising and provides a high level of accuracy in recognizing ASL.
{"title":"An Innovative Method for Automatic American Sign Language Interpretation using Machine Learning and Leap Motion Controller","authors":"Jon Jenkins, S. Rashad","doi":"10.1109/UEMCON53757.2021.9666640","DOIUrl":"https://doi.org/10.1109/UEMCON53757.2021.9666640","url":null,"abstract":"Millions of people globally use some form of sign language in their everyday lives. There is a need for a method of gesture recognition that is as easy to use and ubiquitous as voice recognition is today. In this paper we explore a way to translate from sign language to speech using an innovative method, utilizing the Leap Motion Controller and machine learning algorithms to capture and analyze hand movements in real time, then converting the interpreted signs into spoken word. We seek to build a system that is easy to use, intuitive to understand, adaptable to the individual, and usable in everyday life. This system will be able to work in an adaptive way to learn new signs to expand the dictionary of the system and allow higher accuracy on an individual level. It will have a wide range of applications for healthcare, education, gamification, communication, and more. An optical hand tracking piece of hardware, the Leap Motion Controller will be used to capture hand movements and information to create supervised machine learning models that can be trained to accurately guess American Sign Language (ASL) symbols being signed in real time. Experimental results show that the proposed method is promising and provides a high level of accuracy in recognizing ASL.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123507926","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 : 2021-12-01DOI: 10.1109/UEMCON53757.2021.9666556
J. McShane, Kevin Meehan, Eoghan Furey, M. McAfee
Waste in rivers is an ever-increasing problem. This paper will look at Deep Learning and Computer Vision technologies to determine if they can be applied to the problem domain. Usage of Deep Learning and Computer Vision technologies has grown massively in the last few years thanks to increased computational power, the availability of training data such as ImageNet, and the availability more complex and efficient algorithms. This research investigates two models to determine which one is more suited for the problem domain by evaluating their results based on performing training and testing on a developed waste dataset for the purposes of this research. The dataset is developed four times, each variant incurring more implementation of pre-processing techniques than the other. This resulted in the same dataset being tested four times on both models with varying levels of pre-processing. The first variant of the dataset had no pre-processing, the second with aspect ratio adjusting, the third dataset being augmented by the image data generator, and the fourth by way of an independent augmentation pipeline. The developed waste dataset has images of size 100x100 dimensions regardless of variant. Variant one of the waste datasets contained 1000 images and expanded all the way up to 19,973 images after pipeline augmentation in variant 4. Both VGG-16 and DenseNet-201 will have all four variants implemented on them to investigate which CNN best suits this research domain but also investigate the differences of applying different pre-processing techniques and how this affects results yielded by the two CNN models.
{"title":"Classifying Plastic Waste on River Surfaces utilising CNN and Tensorflow","authors":"J. McShane, Kevin Meehan, Eoghan Furey, M. McAfee","doi":"10.1109/UEMCON53757.2021.9666556","DOIUrl":"https://doi.org/10.1109/UEMCON53757.2021.9666556","url":null,"abstract":"Waste in rivers is an ever-increasing problem. This paper will look at Deep Learning and Computer Vision technologies to determine if they can be applied to the problem domain. Usage of Deep Learning and Computer Vision technologies has grown massively in the last few years thanks to increased computational power, the availability of training data such as ImageNet, and the availability more complex and efficient algorithms. This research investigates two models to determine which one is more suited for the problem domain by evaluating their results based on performing training and testing on a developed waste dataset for the purposes of this research. The dataset is developed four times, each variant incurring more implementation of pre-processing techniques than the other. This resulted in the same dataset being tested four times on both models with varying levels of pre-processing. The first variant of the dataset had no pre-processing, the second with aspect ratio adjusting, the third dataset being augmented by the image data generator, and the fourth by way of an independent augmentation pipeline. The developed waste dataset has images of size 100x100 dimensions regardless of variant. Variant one of the waste datasets contained 1000 images and expanded all the way up to 19,973 images after pipeline augmentation in variant 4. Both VGG-16 and DenseNet-201 will have all four variants implemented on them to investigate which CNN best suits this research domain but also investigate the differences of applying different pre-processing techniques and how this affects results yielded by the two CNN models.","PeriodicalId":127072,"journal":{"name":"2021 IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121305096","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：