Pub Date : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378615
J. Leško, D. Megyesi, Š. Karaffa, M. Hlinková, R. Andoga, R. Bréda
The article is focused on the unmanned aerial vehicle (UAV) Skydog., which is used as a platform for an advanced control system development. A research that was done in area of a UAV control., flight phase classification and sensors., is used to improve mathematical model of the UAV., created in simulation environment Matlab/Simulink. The mathematical model is identical to the real UAV model; hence it can be used as a platform for the advanced system development. The focus of the article is to define concept of the UAV“s integrated system., which consists of advanced control system., advanced flight phase classification system., diagnostics system and improved sensors.
{"title":"A UAV Skydog as a Platform for a Research and a Development of Advanced Control Systems","authors":"J. Leško, D. Megyesi, Š. Karaffa, M. Hlinková, R. Andoga, R. Bréda","doi":"10.1109/SAMI50585.2021.9378615","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378615","url":null,"abstract":"The article is focused on the unmanned aerial vehicle (UAV) Skydog., which is used as a platform for an advanced control system development. A research that was done in area of a UAV control., flight phase classification and sensors., is used to improve mathematical model of the UAV., created in simulation environment Matlab/Simulink. The mathematical model is identical to the real UAV model; hence it can be used as a platform for the advanced system development. The focus of the article is to define concept of the UAV“s integrated system., which consists of advanced control system., advanced flight phase classification system., diagnostics system and improved sensors.","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127096814","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378614
Youssef Bouaziz, E. Royer, Guillaume Bresson, M. Dhome
Appearance changes are a challenge for visual localization in outdoor environments. Revisiting familiar places but retrieving keyframes that were taken under different environmental condition can result in inaccurate localization. To overcome this difficulty, we propose a localization approach able to take advantage of a visual landmark map composed of $N$ sequences gathered at different times and conditions. During this localization process, we exploit information collected in the beginning of the trajectory to compute a ranking function which will be used in the rest of the trajectory to retrieve from the map the keyframes that maximise the number of matched points. The retrieval depends on the geometric distance between the pose of the keyframe and the current pose of the vehicle, and the similarity of this keyframe with the current environmental condition. The results demonstrate that our approach has significantly improved localization performance in challenging conditions (snow, rain, change of season …).
{"title":"Keyframes retrieval for robust long-term visual localization in changing conditions","authors":"Youssef Bouaziz, E. Royer, Guillaume Bresson, M. Dhome","doi":"10.1109/SAMI50585.2021.9378614","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378614","url":null,"abstract":"Appearance changes are a challenge for visual localization in outdoor environments. Revisiting familiar places but retrieving keyframes that were taken under different environmental condition can result in inaccurate localization. To overcome this difficulty, we propose a localization approach able to take advantage of a visual landmark map composed of $N$ sequences gathered at different times and conditions. During this localization process, we exploit information collected in the beginning of the trajectory to compute a ranking function which will be used in the rest of the trajectory to retrieve from the map the keyframes that maximise the number of matched points. The retrieval depends on the geometric distance between the pose of the keyframe and the current pose of the vehicle, and the similarity of this keyframe with the current environmental condition. The results demonstrate that our approach has significantly improved localization performance in challenging conditions (snow, rain, change of season …).","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126935138","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378651
Ján Magyar, P. Sinčák
Personalization is ever more prevalent in digital systems in various application domains. Reinforcement learning is a method often applied to adjust a system's behavior to the user's preferences, but there are a number of hurdles when applying it in this context. We propose a novel neural network architecture for reinforcement learning agents specifically tailored to support personalization - Dynamically Loaded Biases Q-Network. We test our architecture on two environments simulating a personalization task and show that it can simultaneously learn a general behavior and adjust it to different environments.
{"title":"Q-Networks with Dynamically Loaded Biases for Personalization","authors":"Ján Magyar, P. Sinčák","doi":"10.1109/SAMI50585.2021.9378651","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378651","url":null,"abstract":"Personalization is ever more prevalent in digital systems in various application domains. Reinforcement learning is a method often applied to adjust a system's behavior to the user's preferences, but there are a number of hurdles when applying it in this context. We propose a novel neural network architecture for reinforcement learning agents specifically tailored to support personalization - Dynamically Loaded Biases Q-Network. We test our architecture on two environments simulating a personalization task and show that it can simultaneously learn a general behavior and adjust it to different environments.","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"127 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122522917","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378610
Michal Mihálik, M. Hruboš, A. Janota
This work is focused on establishing TCP / IP connection between SICK LD-OEM 1000 laser scanner and PC. Space measurements will be made using this connection. Then we analyze the data and process it. The data is processed by the Monte Carlo Simultaneous localization and mapping algorithm, which will be used to compile a map of the scanned space. This work is based on MATLAB software environment
{"title":"Testing of SLAM methods in the Matlab environment","authors":"Michal Mihálik, M. Hruboš, A. Janota","doi":"10.1109/SAMI50585.2021.9378610","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378610","url":null,"abstract":"This work is focused on establishing TCP / IP connection between SICK LD-OEM 1000 laser scanner and PC. Space measurements will be made using this connection. Then we analyze the data and process it. The data is processed by the Monte Carlo Simultaneous localization and mapping algorithm, which will be used to compile a map of the scanned space. This work is based on MATLAB software environment","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"33 1-2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116732398","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378638
Tran Nguen Bao Ngo, A. Tick
The mushrooming occurrences of cyber criminals in the recent years have provoked an alarm about the drawback of technological growth and the increasing dependence of human beings on technology. The severity of this situation in the business world is even greater and greater than other fields, which leads many people raise a question about the response of external auditors - the ones who are responsible for detecting any accounting faults - towards cybersecurity-attacked companies - the ones which can try their best to hide their difficulties from their investors and stakeholders. Hence, this study investigates whether external auditors pay more attention to cybersecurity-attacked companies by applying higher audit fee charges. Using a sample of 100 global small, medium and large companies, the study has found out that there is a positive relationship between audit fees and breach, which means that external auditors find more risks and exert more efforts when auditing the cybersecurity-attacked companies.
{"title":"External Auditors' Assessments of Cyber-Security Risks","authors":"Tran Nguen Bao Ngo, A. Tick","doi":"10.1109/SAMI50585.2021.9378638","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378638","url":null,"abstract":"The mushrooming occurrences of cyber criminals in the recent years have provoked an alarm about the drawback of technological growth and the increasing dependence of human beings on technology. The severity of this situation in the business world is even greater and greater than other fields, which leads many people raise a question about the response of external auditors - the ones who are responsible for detecting any accounting faults - towards cybersecurity-attacked companies - the ones which can try their best to hide their difficulties from their investors and stakeholders. Hence, this study investigates whether external auditors pay more attention to cybersecurity-attacked companies by applying higher audit fee charges. Using a sample of 100 global small, medium and large companies, the study has found out that there is a positive relationship between audit fees and breach, which means that external auditors find more risks and exert more efforts when auditing the cybersecurity-attacked companies.","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129231434","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378670
Issam Boukhennoufa, X. Zhai, K. Mcdonald-Maier, V. Utti, J. Jackson
An important part of developing a performant assessment algorithm for post-stroke rehabilitation is to achieve a high-precision activity recognition. Convolutional Neural Networks (CNN) are known to give very accurate results, however they require the data to be of a specific structure that differs from the sequential time-series format typically collected from wearable sensors. In this paper, we describe models to improve the activity recognition using the CNN classifier. At first by modifying the Gramian angular field algorithm by encoding all the sensors' channels from a single time window into a single 2D image allows to map the maximum activity characteristics. Feeding the resulting images to a simple 1D CNN classifier improves the accuracy of the test data from 94% for the traditional segmentation approach to 97.06%. Subsequently, we convert the 2D images into the RGB format and use a 2D CNN classifier. This results in increasing the test data accuracy to 97.52%. Finally, we employ transfer learning with the popular VGG_16 model to the RGB images, which yields to improving the accuracy further more to reach 98.53%.
{"title":"Improving the activity recognition using GMAF and transfer learning in post-stroke rehabilitation assessment","authors":"Issam Boukhennoufa, X. Zhai, K. Mcdonald-Maier, V. Utti, J. Jackson","doi":"10.1109/SAMI50585.2021.9378670","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378670","url":null,"abstract":"An important part of developing a performant assessment algorithm for post-stroke rehabilitation is to achieve a high-precision activity recognition. Convolutional Neural Networks (CNN) are known to give very accurate results, however they require the data to be of a specific structure that differs from the sequential time-series format typically collected from wearable sensors. In this paper, we describe models to improve the activity recognition using the CNN classifier. At first by modifying the Gramian angular field algorithm by encoding all the sensors' channels from a single time window into a single 2D image allows to map the maximum activity characteristics. Feeding the resulting images to a simple 1D CNN classifier improves the accuracy of the test data from 94% for the traditional segmentation approach to 97.06%. Subsequently, we convert the 2D images into the RGB format and use a 2D CNN classifier. This results in increasing the test data accuracy to 97.52%. Finally, we employ transfer learning with the popular VGG_16 model to the RGB images, which yields to improving the accuracy further more to reach 98.53%.","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114869255","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378662
Andrea Abela, Thomas Gatt
A serious global waste crisis is currently in effect which originates from our lack of sense of duty. This can be resolved by automating the separation process using AI empowered by weakly supervised learning. A prototype system was created by using pre-trained CNN models in CV such as VGG, ResNet, MobileNet and DenseNet. The prototype showed promising results by having the best algorithms obtain an F1-score of over 80% on 2 datasets known as TrashNet and MINC. Some algorithms were also quite efficient, reaching over 10FPS while maintaining less than 10Mb. The localisation accuracy generated from the CAMs of the best models has shown to be around 83% on TrashNet and around 69% on MINC. These results show that not only is it possible through AI to accurately and efficiently classify waste through datasets, but it can also be used to integrate accurate localisation via weak supervision for easier data annotation.
{"title":"Using Class Activation Maps on Deep Neural Networks to Localise Waste Classifications","authors":"Andrea Abela, Thomas Gatt","doi":"10.1109/SAMI50585.2021.9378662","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378662","url":null,"abstract":"A serious global waste crisis is currently in effect which originates from our lack of sense of duty. This can be resolved by automating the separation process using AI empowered by weakly supervised learning. A prototype system was created by using pre-trained CNN models in CV such as VGG, ResNet, MobileNet and DenseNet. The prototype showed promising results by having the best algorithms obtain an F1-score of over 80% on 2 datasets known as TrashNet and MINC. Some algorithms were also quite efficient, reaching over 10FPS while maintaining less than 10Mb. The localisation accuracy generated from the CAMs of the best models has shown to be around 83% on TrashNet and around 69% on MINC. These results show that not only is it possible through AI to accurately and efficiently classify waste through datasets, but it can also be used to integrate accurate localisation via weak supervision for easier data annotation.","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126342847","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378686
Ivan Cík, Andrindrasana David Rasamoelina, M. Mach, P. Sinčák
Machine learning has become an integral part of technology in today's world. The field of artificial intelligence is the subject of research by a wide scientific community. In particular, through improved methodology, the availability of big data, and increased computing power, today's machine learning algorithms can achieve excellent performance that sometimes even exceeds the human level. However, due to their nested nonlinear structure, these models are generally considered to be “Black boxes” that do not provide any information about what exactly leads them to provide a specific output. This raised the need to interpret these algorithms and understand how they work as they are applied even in areas where they can cause critical damage. This article describes Integrated Gradients [1] and Layer-wise Relevance Propagation [2] methods and presents individual experiments with. In experiments we have used well-known datasets like MNIST[3], MNIST-Fashion dataset[4], Imagenette and Imagewoof which are subsets of ImageNet [5].
{"title":"Explaining Deep Neural Network using Layer-wise Relevance Propagation and Integrated Gradients","authors":"Ivan Cík, Andrindrasana David Rasamoelina, M. Mach, P. Sinčák","doi":"10.1109/SAMI50585.2021.9378686","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378686","url":null,"abstract":"Machine learning has become an integral part of technology in today's world. The field of artificial intelligence is the subject of research by a wide scientific community. In particular, through improved methodology, the availability of big data, and increased computing power, today's machine learning algorithms can achieve excellent performance that sometimes even exceeds the human level. However, due to their nested nonlinear structure, these models are generally considered to be “Black boxes” that do not provide any information about what exactly leads them to provide a specific output. This raised the need to interpret these algorithms and understand how they work as they are applied even in areas where they can cause critical damage. This article describes Integrated Gradients [1] and Layer-wise Relevance Propagation [2] methods and presents individual experiments with. In experiments we have used well-known datasets like MNIST[3], MNIST-Fashion dataset[4], Imagenette and Imagewoof which are subsets of ImageNet [5].","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125595586","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378643
P. Peniak, E. Bubeníková, A. Kanáliková
In the paper, we share our research results and experiences with the industrial integration of IoT devices and OPC UA systems, assuming it is one of major enablers for Industry 4.0 use-cases. The main focus is paid on the numerical model of integration gateway, which could enable integration of various IoT devices, embedded and constrained systems, where implementation of full-scale OPC UA protocol is not possible. The goal is to create the appropriate model for OPC UA/IoT device integrations, verify and test it with implemented integration gateway. The gateway implementation scope and testing was concentrated on OPC DA and two major IoT protocols: MQTT and COAP.
{"title":"Extended gateway model for OPC UA/IoT device integration","authors":"P. Peniak, E. Bubeníková, A. Kanáliková","doi":"10.1109/SAMI50585.2021.9378643","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378643","url":null,"abstract":"In the paper, we share our research results and experiences with the industrial integration of IoT devices and OPC UA systems, assuming it is one of major enablers for Industry 4.0 use-cases. The main focus is paid on the numerical model of integration gateway, which could enable integration of various IoT devices, embedded and constrained systems, where implementation of full-scale OPC UA protocol is not possible. The goal is to create the appropriate model for OPC UA/IoT device integrations, verify and test it with implemented integration gateway. The gateway implementation scope and testing was concentrated on OPC DA and two major IoT protocols: MQTT and COAP.","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130802948","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 : 2021-01-21DOI: 10.1109/SAMI50585.2021.9378642
G. Varga, Andras Kondakor, Márton Antal
One of the many scenarios associated with self-driving cars is related to automatic parking assistance systems. In our research, we implemented an Autonomous Valet Parking System and built a simulated, dynamically changing environment to test the method. This paper presents the three main parts of the system namely Parking Space Detection using ultrasonic sensors and LIDARs, Path Planning, which is achieved with Hybrid A * and RTR+CCRS planners, and Dynamic Object Detection employing neural networks and image segmentation on the output of an RGB-D camera.
{"title":"Developing an Autonomous Valet Parking System in Simulated Environment","authors":"G. Varga, Andras Kondakor, Márton Antal","doi":"10.1109/SAMI50585.2021.9378642","DOIUrl":"https://doi.org/10.1109/SAMI50585.2021.9378642","url":null,"abstract":"One of the many scenarios associated with self-driving cars is related to automatic parking assistance systems. In our research, we implemented an Autonomous Valet Parking System and built a simulated, dynamically changing environment to test the method. This paper presents the three main parts of the system namely Parking Space Detection using ultrasonic sensors and LIDARs, Path Planning, which is achieved with Hybrid A * and RTR+CCRS planners, and Dynamic Object Detection employing neural networks and image segmentation on the output of an RGB-D camera.","PeriodicalId":402414,"journal":{"name":"2021 IEEE 19th World Symposium on Applied Machine Intelligence and Informatics (SAMI)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124418911","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：