Pub Date : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029428
Catalin Vladu, L. Prodan, A. Iovanovici
This paper addresses the automated identification of violent acts from CCTV video streams using a Deep Learning model under constrained resources. While this process typically involves a powerful setup, it is useful to accelerate the training and get accurate results using more modest computational resources that would bring automatic recognition of violent acts closer to common surveillance resources. Our results provide 94.98% accuracy, on par with the state-of-the-art, but at a fraction of the training time. This translates into lower energy requirements and allows a broader deployment on large scale (urban) autonomous surveillance networks while providing an increased privacy towards citizens and lower chances of abuse from authorities.
{"title":"Resource Constrained, Fast Convergence Training for Violence Detection in Video Streams","authors":"Catalin Vladu, L. Prodan, A. Iovanovici","doi":"10.1109/CINTI-MACRo57952.2022.10029428","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029428","url":null,"abstract":"This paper addresses the automated identification of violent acts from CCTV video streams using a Deep Learning model under constrained resources. While this process typically involves a powerful setup, it is useful to accelerate the training and get accurate results using more modest computational resources that would bring automatic recognition of violent acts closer to common surveillance resources. Our results provide 94.98% accuracy, on par with the state-of-the-art, but at a fraction of the training time. This translates into lower energy requirements and allows a broader deployment on large scale (urban) autonomous surveillance networks while providing an increased privacy towards citizens and lower chances of abuse from authorities.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"6 1","pages":"000239-000244"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79936401","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029569
M. Ermakov, Ilya Makarov
Brand logo recognition can be viewed as identification and classification task. It finds many usages such as market discovery, target advertising, etc. The number of logos growth every year and logo itself can appear in vast variety of contexts, therefore we propose a two-step few-shot framework. We describe a novel combination of universal logo detector and few-shot classifier. The logo detector is based on YOLOv5 and is used to find the areas on the image where logos are located. With this state-of-the-art single-stage object detector we achieved higher precision than similar double-stage solutions. To classify detected logos we propose few-shot classifier which consists of ensemble of pretrained feature extractors and fine-tuned head.
{"title":"Few-shot Logo Recognition in the Wild","authors":"M. Ermakov, Ilya Makarov","doi":"10.1109/CINTI-MACRo57952.2022.10029569","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029569","url":null,"abstract":"Brand logo recognition can be viewed as identification and classification task. It finds many usages such as market discovery, target advertising, etc. The number of logos growth every year and logo itself can appear in vast variety of contexts, therefore we propose a two-step few-shot framework. We describe a novel combination of universal logo detector and few-shot classifier. The logo detector is based on YOLOv5 and is used to find the areas on the image where logos are located. With this state-of-the-art single-stage object detector we achieved higher precision than similar double-stage solutions. To classify detected logos we propose few-shot classifier which consists of ensemble of pretrained feature extractors and fine-tuned head.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"68 1","pages":"000393-000398"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90367705","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029588
Grega Vrbancic, Rok Kukovec, V. Podgorelec, S. Salcedo-Sanz, Iztok Fister
Smart agriculture takes advantage of modern computational approaches that vary from IoT, cloud computing, and artificial intelligence. The primary aim is to assist the farming process. Pest detection is one of the objectives within the area of smart agriculture. It is mainly solved by computer vision approaches, usually combined with machine learning (ML) algorithms. In this paper, we propose a solution for detecting Arion rufus snails that have emerged in Central Europe and are one of the most prolific threats to agriculture in that place. Practical experiments reveal that our method is helpful in this real-world application and opens several future challenges and lines of research.
{"title":"Profiling the Arion rufus snails with computer vision","authors":"Grega Vrbancic, Rok Kukovec, V. Podgorelec, S. Salcedo-Sanz, Iztok Fister","doi":"10.1109/CINTI-MACRo57952.2022.10029588","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029588","url":null,"abstract":"Smart agriculture takes advantage of modern computational approaches that vary from IoT, cloud computing, and artificial intelligence. The primary aim is to assist the farming process. Pest detection is one of the objectives within the area of smart agriculture. It is mainly solved by computer vision approaches, usually combined with machine learning (ML) algorithms. In this paper, we propose a solution for detecting Arion rufus snails that have emerged in Central Europe and are one of the most prolific threats to agriculture in that place. Practical experiments reveal that our method is helpful in this real-world application and opens several future challenges and lines of research.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"9 1","pages":"000369-000374"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90776038","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029465
Adrienn Deak, Z. Szántó, Áron Fehér, L. Márton
This work presents a solution for distant control of industrial robots using mobile devices. The developed application can monitor the motion of the robot based on video and sensor information, it can send commands and control scripts to the robot, and it ensures the jogging–like manual control of the robot. In addition to the aforementioned functionalities, the application can analyze the teleoperation performances of the user. A set of performance metrics were introduced to rate user performances. They are useful to evaluate and educate remote manual robot control techniques. Experimental measurements are also presented to show the applicability of the developed remote control application and user performance evaluation method.
{"title":"Smartphone–controlled industrial robots: Design and user performance evaluation","authors":"Adrienn Deak, Z. Szántó, Áron Fehér, L. Márton","doi":"10.1109/CINTI-MACRo57952.2022.10029465","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029465","url":null,"abstract":"This work presents a solution for distant control of industrial robots using mobile devices. The developed application can monitor the motion of the robot based on video and sensor information, it can send commands and control scripts to the robot, and it ensures the jogging–like manual control of the robot. In addition to the aforementioned functionalities, the application can analyze the teleoperation performances of the user. A set of performance metrics were introduced to rate user performances. They are useful to evaluate and educate remote manual robot control techniques. Experimental measurements are also presented to show the applicability of the developed remote control application and user performance evaluation method.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"1 1","pages":"000083-000088"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89724420","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029407
Anett Popovics, V. Szekeres
Today, successful companies are all placing a strong emphasis on corporate social responsibility: they are involved in activities and support causes that serve the interests of a community. The emergence of a group of responsible consumers’ is not new, who place great emphasis on buying environmentally responsible and ethical products and who also value transparent communication from companies. In our quantitative methodological research, we investigated whether there is a correlation between social activism and the purchase of ethical products, and which consumer groups are active buyers of ethical products. Our results confirmed both our hypotheses: we have shown that this attitude is reflected in consumer behaviour and could contribute to increasing social responsibility activity in the future.
{"title":"Analysis of social responsibility and consumer activity through primary research","authors":"Anett Popovics, V. Szekeres","doi":"10.1109/CINTI-MACRo57952.2022.10029407","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029407","url":null,"abstract":"Today, successful companies are all placing a strong emphasis on corporate social responsibility: they are involved in activities and support causes that serve the interests of a community. The emergence of a group of responsible consumers’ is not new, who place great emphasis on buying environmentally responsible and ethical products and who also value transparent communication from companies. In our quantitative methodological research, we investigated whether there is a correlation between social activism and the purchase of ethical products, and which consumer groups are active buyers of ethical products. Our results confirmed both our hypotheses: we have shown that this attitude is reflected in consumer behaviour and could contribute to increasing social responsibility activity in the future.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"7 1","pages":"000163-000166"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87499451","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029403
Saba Danaei, Arsam Bostani, Seyed Vahid Moravvej, F. Mohammadi, R. Alizadehsani, A. Shoeibi, H. Alinejad-Rokny, Saeid Nahavandi
Myocarditis occurs when the heart muscle becomes inflamed and inflammation occurs when your body’s immune system responds to infections. It can be diagnosed using cardiac magnetic resonance image (MRI), a non-invasive imaging technique with the possibility of operator bias. This paper proposes a hybrid method of deep reinforcement learning-based algorithms and meta-heuristics algorithms. A mutual learning-based artificial bee colony (ML-ABC) is employed for initial weight, which adjusts the candidate food source generated with the higher fitness between two individuals determined by a mutual learning factor. Moreover, a sequential decision-making process investigates the imbalanced classification issue, in which a convolutional neural network (CNN) is used as the foundation for policy architecture. At first, initial weights are produced using the ML-ABC algorithm. After that, the agent receives a sample at each phase and classifies it, obtaining environmental rewards. The minority class receives more rewards than the majority class. Eventually, the agent discovers an ideal strategy with the aid of a specific reward function and a beneficial learning environment. We evaluate our proposed approach on the Z-Alizadeh Sani myocarditis dataset based on standard criteria and demonstrate that the proposed method gives superior myocarditis diagnosis performance.
{"title":"Myocarditis Diagnosis: A Method using Mutual Learning-Based ABC and Reinforcement Learning","authors":"Saba Danaei, Arsam Bostani, Seyed Vahid Moravvej, F. Mohammadi, R. Alizadehsani, A. Shoeibi, H. Alinejad-Rokny, Saeid Nahavandi","doi":"10.1109/CINTI-MACRo57952.2022.10029403","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029403","url":null,"abstract":"Myocarditis occurs when the heart muscle becomes inflamed and inflammation occurs when your body’s immune system responds to infections. It can be diagnosed using cardiac magnetic resonance image (MRI), a non-invasive imaging technique with the possibility of operator bias. This paper proposes a hybrid method of deep reinforcement learning-based algorithms and meta-heuristics algorithms. A mutual learning-based artificial bee colony (ML-ABC) is employed for initial weight, which adjusts the candidate food source generated with the higher fitness between two individuals determined by a mutual learning factor. Moreover, a sequential decision-making process investigates the imbalanced classification issue, in which a convolutional neural network (CNN) is used as the foundation for policy architecture. At first, initial weights are produced using the ML-ABC algorithm. After that, the agent receives a sample at each phase and classifies it, obtaining environmental rewards. The minority class receives more rewards than the majority class. Eventually, the agent discovers an ideal strategy with the aid of a specific reward function and a beneficial learning environment. We evaluate our proposed approach on the Z-Alizadeh Sani myocarditis dataset based on standard criteria and demonstrate that the proposed method gives superior myocarditis diagnosis performance.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"30 1","pages":"000265-000270"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88509853","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029564
A. I. Károly, Sebestyen Tirczka, Tamas Piricz, P. Galambos
Digital pathology has many advantages, so the need for digitizing already existing archives naturally arises. However, the fact that there is no standard way of storing pathology archives makes it difficult to provide an automated solution. In this paper, we tackle this problem with a robotic system, which uses a deep convolutional neural network and traditional image processing methods to automatically detect and localize the pathology samples and perform pick and place to organize the samples in a rack that can be directly inserted into the whole slide imaging (WSI) scanner. We were able to achieve a 90% success rate for the pick and place process. This paper introduces the hardware setup and software components that we used for our system and briefly explains the detection procedure.
{"title":"Robotic Manipulation of Pathological Slides Powered by Deep Learning and Classical Image Processing","authors":"A. I. Károly, Sebestyen Tirczka, Tamas Piricz, P. Galambos","doi":"10.1109/CINTI-MACRo57952.2022.10029564","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029564","url":null,"abstract":"Digital pathology has many advantages, so the need for digitizing already existing archives naturally arises. However, the fact that there is no standard way of storing pathology archives makes it difficult to provide an automated solution. In this paper, we tackle this problem with a robotic system, which uses a deep convolutional neural network and traditional image processing methods to automatically detect and localize the pathology samples and perform pick and place to organize the samples in a rack that can be directly inserted into the whole slide imaging (WSI) scanner. We were able to achieve a 90% success rate for the pick and place process. This paper introduces the hardware setup and software components that we used for our system and briefly explains the detection procedure.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"34 1","pages":"000387-000392"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83618899","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029438
I. M. Putrama, P. Martinek
There are still few research methods proposed to convert relational databases to graph databases. Although a graph database has been equipped with a scripting language to use for querying and converting the data, it still requires time-consuming efforts by the domain expert to analyze the various constraints present in the source database. This paper proposes a novel technique to help automate the conversion by extracting relational database metadata and then sorting the entity relationships before converting them into graphs. To validate the conversion results, the total number of records in the source database with the number of nodes and edges created in the graph database are compared, and the node properties are validated for consistency using a probabilistic data structure. Based on our test results, their completeness can be checked accurately and efficiently with test parameters that can be adjusted according to the size of the source database.
{"title":"An Automated Graph Construction Approach from Relational Databases to Neo4j","authors":"I. M. Putrama, P. Martinek","doi":"10.1109/CINTI-MACRo57952.2022.10029438","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029438","url":null,"abstract":"There are still few research methods proposed to convert relational databases to graph databases. Although a graph database has been equipped with a scripting language to use for querying and converting the data, it still requires time-consuming efforts by the domain expert to analyze the various constraints present in the source database. This paper proposes a novel technique to help automate the conversion by extracting relational database metadata and then sorting the entity relationships before converting them into graphs. To validate the conversion results, the total number of records in the source database with the number of nodes and edges created in the graph database are compared, and the node properties are validated for consistency using a probabilistic data structure. Based on our test results, their completeness can be checked accurately and efficiently with test parameters that can be adjusted according to the size of the source database.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"64 6","pages":"000131-000136"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91498760","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 : 2022-11-21DOI: 10.1109/CINTI-MACRo57952.2022.10029502
Michal Kolárik, M. Sarnovský, Ján Paralič, P. Butka
The ability to explain the reasons for one’s decisions to others is an important aspect of being human intelligence. We will look at the explainability aspects of the deep learning models, which are most frequently used in medical image processing tasks. The Explainability of machine learning models in medicine is essential for understanding how the particular ML model works and how it solves the problems it was designed for. The work presented in this paper focuses on the classification of lung CT scans for the detection of COVID-19 patients. We used CNN and DenseNet models for the classification and explored the application of selected visual explainability techniques to provide insight into how the model works when processing the images.
{"title":"Explainability of deep learning models in medical image classification","authors":"Michal Kolárik, M. Sarnovský, Ján Paralič, P. Butka","doi":"10.1109/CINTI-MACRo57952.2022.10029502","DOIUrl":"https://doi.org/10.1109/CINTI-MACRo57952.2022.10029502","url":null,"abstract":"The ability to explain the reasons for one’s decisions to others is an important aspect of being human intelligence. We will look at the explainability aspects of the deep learning models, which are most frequently used in medical image processing tasks. The Explainability of machine learning models in medicine is essential for understanding how the particular ML model works and how it solves the problems it was designed for. The work presented in this paper focuses on the classification of lung CT scans for the detection of COVID-19 patients. We used CNN and DenseNet models for the classification and explored the application of selected visual explainability techniques to provide insight into how the model works when processing the images.","PeriodicalId":18535,"journal":{"name":"Micro","volume":"11 1","pages":"000233-000238"},"PeriodicalIF":0.0,"publicationDate":"2022-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76157517","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：