As a basic visual recognition problem in computer vision, object detection has made great progress based on traditional manual features and deep learning algorithms. However, researches on small object detection ha ve only begun to appear in recent years, which has become a hot and difficult point in the field and most of them are improved on the basis of existing object detection algorithms to enhance the detection accuracy. With the rapid development of deep learning, small object detection based on deep learning has made great progress, which has wide application requirements in the fields of automatic driving, remote sensing image detection, criminal investigation and other fields, so the research on small object detection has strong practical values. In this paper, the existing research on small target detection is reviewed in detail. Firstly, the existing algorithms are divided into one stage and two stages according to the number of detection stages, and then the characteristics of these algorithms are analyzed; Secondly, the small object detection datasets commonly used are introduced. Finally, the challenges of small object detection are summarized, and the future research directions are prospected.
{"title":"A Survey of Small Object Detection Based on Deep Learning","authors":"Zhenghua Zhang, Jiang Ling, Qingqing Hong","doi":"10.12792/icisip2021.009","DOIUrl":"https://doi.org/10.12792/icisip2021.009","url":null,"abstract":"As a basic visual recognition problem in computer vision, object detection has made great progress based on traditional manual features and deep learning algorithms. However, researches on small object detection ha ve only begun to appear in recent years, which has become a hot and difficult point in the field and most of them are improved on the basis of existing object detection algorithms to enhance the detection accuracy. With the rapid development of deep learning, small object detection based on deep learning has made great progress, which has wide application requirements in the fields of automatic driving, remote sensing image detection, criminal investigation and other fields, so the research on small object detection has strong practical values. In this paper, the existing research on small target detection is reviewed in detail. Firstly, the existing algorithms are divided into one stage and two stages according to the number of detection stages, and then the characteristics of these algorithms are analyzed; Secondly, the small object detection datasets commonly used are introduced. Finally, the challenges of small object detection are summarized, and the future research directions are prospected.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116976727","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}
Rice is one of the main food crops in China, and rice diseases have become an important factor influencing the increase in food production losses in China. Traditional manual identification of rice diseases is time-consuming and labor-intensive. Machine learning algorithms have improved this problem and have been applied to the field of smart agriculture. The convolutional neural network (CNN) in deep learning has a significant effect on rice disease recognition relying on the characteristics of automatically extracting features. Aiming at five major rice diseases such as sheath blight, rice blast, bacterial leaf blight, rice smut and brown spot, this paper proposed a rice disease identification system using lightweight MobileNetV2. The identification results are uploaded and saved to the cloud database. Based on the lightweight model MobileNetV2, the system uses the channel pruning method to further compress the model. Compared with the original model, the memory usage has been reduced by 74%, the number of floating-point operations per second (FLOPS) has been reduced by 49%, the number of parameters has been reduced by 50%, and the accuracy of rice disease identification has increased by 0.16% to 90.84%.
{"title":"Rice Disease Identification System Using Lightweight MobileNetV2","authors":"Zhenghua Zhang, Yifeng Gu, Qingqing Hong","doi":"10.12792/icisip2021.007","DOIUrl":"https://doi.org/10.12792/icisip2021.007","url":null,"abstract":"Rice is one of the main food crops in China, and rice diseases have become an important factor influencing the increase in food production losses in China. Traditional manual identification of rice diseases is time-consuming and labor-intensive. Machine learning algorithms have improved this problem and have been applied to the field of smart agriculture. The convolutional neural network (CNN) in deep learning has a significant effect on rice disease recognition relying on the characteristics of automatically extracting features. Aiming at five major rice diseases such as sheath blight, rice blast, bacterial leaf blight, rice smut and brown spot, this paper proposed a rice disease identification system using lightweight MobileNetV2. The identification results are uploaded and saved to the cloud database. Based on the lightweight model MobileNetV2, the system uses the channel pruning method to further compress the model. Compared with the original model, the memory usage has been reduced by 74%, the number of floating-point operations per second (FLOPS) has been reduced by 49%, the number of parameters has been reduced by 50%, and the accuracy of rice disease identification has increased by 0.16% to 90.84%.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124753389","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}
Zhen Li, Yuren Du, Qingqing Hong, S. Serikawa, Lifeng Zhang
{"title":"A 3D Object Detection Framework for Intelligent Driving using YOLOv4","authors":"Zhen Li, Yuren Du, Qingqing Hong, S. Serikawa, Lifeng Zhang","doi":"10.12792/icisip2021.004","DOIUrl":"https://doi.org/10.12792/icisip2021.004","url":null,"abstract":"","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"178 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131536296","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}
AGVs (Automated Guided Vehicles) are widely used in factories and warehouses. Functions such as position estimation are indispensable for unmanned transport robots. We have developed a new method to estimate the position of an object using a cross laser and a camera. In this study, we verified the measurement accuracy by constructing the estimation principle and implementing the system. The results of position estimation at arbitrary measurement points showed that the measurement error was small, averaging 26 mm and maximum 49 mm, confirming that our system can estimate the position accurately.
{"title":"Method for Estimation of Position of an Object in a Room Using Cross-Line Laser","authors":"Masaya Okamoto, Shiyuan Yang, S. Serikawa","doi":"10.12792/icisip2021.019","DOIUrl":"https://doi.org/10.12792/icisip2021.019","url":null,"abstract":"AGVs (Automated Guided Vehicles) are widely used in factories and warehouses. Functions such as position estimation are indispensable for unmanned transport robots. We have developed a new method to estimate the position of an object using a cross laser and a camera. In this study, we verified the measurement accuracy by constructing the estimation principle and implementing the system. The results of position estimation at arbitrary measurement points showed that the measurement error was small, averaging 26 mm and maximum 49 mm, confirming that our system can estimate the position accurately.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121053613","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}
Sensor-based human activity recognition (HAR) has gained its momentum and become an active research topic due to the advance of machine learning (ML) algorithms and ubiquitous sensing devices in our daily life. Recent research trend in ML algorithms for HAR is deep learning-based approaches that have already developed state-of-the-art learning models in various tasks. However, complex deep learning models may not be the best choice when it comes to data sufficiency problems and model transparency. Exploratory data analysis (EDA) can benefit feature extraction, which is an important step in a machine learning pipeline. In this study, to explore sensor-based HAR, a widely used HAR dataset is adopted to examine the effectiveness of time series feature extraction together with conventional machine learning models. Experimental results show that EDA can be beneficial for obtaining data insights and determining better features for HAR classification.
{"title":"Exploration of Sensor-Based Activity Recognition Based on Time Series Feature Extraction","authors":"Wen-Hui Chen, Ting Chen, Cheng-Han Tsai","doi":"10.12792/icisip2021.023","DOIUrl":"https://doi.org/10.12792/icisip2021.023","url":null,"abstract":"Sensor-based human activity recognition (HAR) has gained its momentum and become an active research topic due to the advance of machine learning (ML) algorithms and ubiquitous sensing devices in our daily life. Recent research trend in ML algorithms for HAR is deep learning-based approaches that have already developed state-of-the-art learning models in various tasks. However, complex deep learning models may not be the best choice when it comes to data sufficiency problems and model transparency. Exploratory data analysis (EDA) can benefit feature extraction, which is an important step in a machine learning pipeline. In this study, to explore sensor-based HAR, a widely used HAR dataset is adopted to examine the effectiveness of time series feature extraction together with conventional machine learning models. Experimental results show that EDA can be beneficial for obtaining data insights and determining better features for HAR classification.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129600065","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}
There are three output types of sensors, which represent a voltage output type sensor, a current output type sensor, and a resistance change type sensor. If the output type is different, the detection circuit is also different. In the previous study, one unique analog detection circuit that can measure the output of different sensors has been proposed. However, the circuit had a switch and had to be switched manually. Therefore, the user needed to know in advance what type of output the sensor had. In this study, the switch is automatically switched according to the sensor type. We classify sensors into energy conversion type (voltage output type and current output type) and energy control type (resistance change type), and propose a method to automatically identify them. Using three types of sensors, we experimentally investigated whether they could be identified correctly. As a result, it became clear that any sensor can be automatically identified. For this reason, we do not need to know the type of sensor in advance. The switch is automatically switched according to the type of sensor. This makes it possible to operate the sensor correctly simply by connecting the sensor to one circuit.
{"title":"Proposal of a Method to Automatically Identify a Sensor as Energy Conversion or Energy Control","authors":"Kei Sato, S. Serikawa","doi":"10.12792/icisip2021.029","DOIUrl":"https://doi.org/10.12792/icisip2021.029","url":null,"abstract":"There are three output types of sensors, which represent a voltage output type sensor, a current output type sensor, and a resistance change type sensor. If the output type is different, the detection circuit is also different. In the previous study, one unique analog detection circuit that can measure the output of different sensors has been proposed. However, the circuit had a switch and had to be switched manually. Therefore, the user needed to know in advance what type of output the sensor had. In this study, the switch is automatically switched according to the sensor type. We classify sensors into energy conversion type (voltage output type and current output type) and energy control type (resistance change type), and propose a method to automatically identify them. Using three types of sensors, we experimentally investigated whether they could be identified correctly. As a result, it became clear that any sensor can be automatically identified. For this reason, we do not need to know the type of sensor in advance. The switch is automatically switched according to the type of sensor. This makes it possible to operate the sensor correctly simply by connecting the sensor to one circuit.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"100 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129322364","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}
This paper describes the development of a predictive maintenance system for cutting machines. In recent years, IoT and AI systems have been developed actively. As a result, sensors and embedded systems are becoming cheaper. Small and medium-sized companies attempt to use these inexpensive embedded systems for predictive maintenance. Therefore, we are developing the AI predictive maintenance system for these companies. In the system, the cutting sound emitted by a cutting machine is acquired by a sensor and an embedded system. The differences in the sounds are analyzed by AI using MATLAB and TensorFlow to predict the wear and tear of the tip of blade. The system was able to predict the tip wear degree with 90.5% accuracy.
{"title":"A Development of AI Predictive Maintenance System using IoT Sensing","authors":"K. Hayakawa, A. Heima, M. Ozaki, Satoshi Yoshida","doi":"10.12792/icisip2021.025","DOIUrl":"https://doi.org/10.12792/icisip2021.025","url":null,"abstract":"This paper describes the development of a predictive maintenance system for cutting machines. In recent years, IoT and AI systems have been developed actively. As a result, sensors and embedded systems are becoming cheaper. Small and medium-sized companies attempt to use these inexpensive embedded systems for predictive maintenance. Therefore, we are developing the AI predictive maintenance system for these companies. In the system, the cutting sound emitted by a cutting machine is acquired by a sensor and an embedded system. The differences in the sounds are analyzed by AI using MATLAB and TensorFlow to predict the wear and tear of the tip of blade. The system was able to predict the tip wear degree with 90.5% accuracy.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127384236","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}
It is difficult for elderly people and people with physical disabilities to open and close windows and curtains whenever the environment changes. Therefore, we propose an automatic window and curtain opening/closing system that uses fuzzy technology to automatically adjust the environment to make people feel comfortable. The system automatically opens and closes the windows and curtains based on the criteria of discomfort or comfort for the following items: illumination, temperature, carbon dioxide concentration, noise, and wind speed. Windows are judged based on the above five criteria, and curtains are judged based on four criteria, excluding carbon dioxide concentration.
{"title":"Automatic Opening and Closing System for Windows and Curtains Using Fuzzy","authors":"Toshimasa Noda, Yuhki Kitazono","doi":"10.12792/icisip2021.012","DOIUrl":"https://doi.org/10.12792/icisip2021.012","url":null,"abstract":"It is difficult for elderly people and people with physical disabilities to open and close windows and curtains whenever the environment changes. Therefore, we propose an automatic window and curtain opening/closing system that uses fuzzy technology to automatically adjust the environment to make people feel comfortable. The system automatically opens and closes the windows and curtains based on the criteria of discomfort or comfort for the following items: illumination, temperature, carbon dioxide concentration, noise, and wind speed. Windows are judged based on the above five criteria, and curtains are judged based on four criteria, excluding carbon dioxide concentration.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129940343","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}
Satoshi Watanabe, Takahiro Sugiyama, Naofumi Nakaya, N. Shirahama
To investigate the effects of listening to music, analyses of heart rate variability (HRV) by discrete Fourier transform (DFT) have been widely conducted. These methods are useful to estimate the autonomic nervous system activity (sympathetic or parasympathetic activity). And to examine the effects of listening to “ Dengaku ” music, one experiment is carried out. Eleven healthy Japanese people participate in the experiment, and one piece of “ Dengaku ” music is employed as the test piece. All participants are asked to listen to the test piece individually, and their HRV are recorded and analyzed by DFT. The experiment result shows that the sympathetic nerve activities of all participants tend to decrease when they are listening to the test piece. This fact is a rare case (usually, there are some participants whose sympathetic nerve activities tend to increase when they are listening to music). , variation of LF/HF indexes are reduced) are obtained when participants are listening to “ Dengaku ” music. However, to obtain reliable conclusions, conducting more investigations would be especially important.
{"title":"Effects of Listening of \"Dengaku\" Music (One of Japanese Traditional Music) to the Changes of Heart Rate Variability","authors":"Satoshi Watanabe, Takahiro Sugiyama, Naofumi Nakaya, N. Shirahama","doi":"10.12792/icisip2021.016","DOIUrl":"https://doi.org/10.12792/icisip2021.016","url":null,"abstract":"To investigate the effects of listening to music, analyses of heart rate variability (HRV) by discrete Fourier transform (DFT) have been widely conducted. These methods are useful to estimate the autonomic nervous system activity (sympathetic or parasympathetic activity). And to examine the effects of listening to “ Dengaku ” music, one experiment is carried out. Eleven healthy Japanese people participate in the experiment, and one piece of “ Dengaku ” music is employed as the test piece. All participants are asked to listen to the test piece individually, and their HRV are recorded and analyzed by DFT. The experiment result shows that the sympathetic nerve activities of all participants tend to decrease when they are listening to the test piece. This fact is a rare case (usually, there are some participants whose sympathetic nerve activities tend to increase when they are listening to music). , variation of LF/HF indexes are reduced) are obtained when participants are listening to “ Dengaku ” music. However, to obtain reliable conclusions, conducting more investigations would be especially important.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124135562","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}
Yudai Kubo, Hidemitsu Arimura, Shenglin Mu, S. Nishifuji, Shota Nakashima
Currently, the number of elderly people in the world is increasing. As a result, the number of accidents involving elderly people falling when movement is increasing. Development of mobility support systems is necessary for them to move safely. Therefore, the system was developed to help wheelchairs move by identifying the type of road surface in front of them. The system used ultrasonic sensors attached to the wheelchair to identify the road surface. Then, the method for identifying four types of road surfaces using Support Vector Machines (SVM) was proposed for the road surface identification method that constitutes the mobility support system. However, in the previous study, only the case where measured road surface didn't change was verified. This made it impossible to make early identification when the road surface changed during measurement. In this paper, the new road surface identification method using ultrasonic sensors is proposed. The proposed method makes it possible to identify the boundary of a road surface when it changes. In addition, the method improves the early detection performance. In order to verify the performance of early identification road boundary, two road surfaces with different roughness were measured in succession. As a result, the proposed method was able to identify at before entering the road boundary. This confirms the effectiveness of the road surface identification method that takes the time series into account for sample obtainment.
{"title":"A Road Surface Identification Method Improved Early Detection Performance Using Ultrasonic Sensors","authors":"Yudai Kubo, Hidemitsu Arimura, Shenglin Mu, S. Nishifuji, Shota Nakashima","doi":"10.12792/icisip2021.021","DOIUrl":"https://doi.org/10.12792/icisip2021.021","url":null,"abstract":"Currently, the number of elderly people in the world is increasing. As a result, the number of accidents involving elderly people falling when movement is increasing. Development of mobility support systems is necessary for them to move safely. Therefore, the system was developed to help wheelchairs move by identifying the type of road surface in front of them. The system used ultrasonic sensors attached to the wheelchair to identify the road surface. Then, the method for identifying four types of road surfaces using Support Vector Machines (SVM) was proposed for the road surface identification method that constitutes the mobility support system. However, in the previous study, only the case where measured road surface didn't change was verified. This made it impossible to make early identification when the road surface changed during measurement. In this paper, the new road surface identification method using ultrasonic sensors is proposed. The proposed method makes it possible to identify the boundary of a road surface when it changes. In addition, the method improves the early detection performance. In order to verify the performance of early identification road boundary, two road surfaces with different roughness were measured in succession. As a result, the proposed method was able to identify at before entering the road boundary. This confirms the effectiveness of the road surface identification method that takes the time series into account for sample obtainment.","PeriodicalId":431446,"journal":{"name":"The Proceedings of The 8th International Conference on Intelligent Systems and Image Processing 2021","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126962460","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：