List of explanations
Deep Learning
Deep Learning
Deep Learning
Examples of research themes
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Examples of research themes
Research Achievements (Excerpts)
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Deep Learning
List of explanations
Deep learning generally refers to a deep neural network and is a type of machine learning algorithm. It is used to estimate what is in an image or other specific information, or to predict the future from past time series. Our laboratory conducts joint research with various external institutions and is working on real-world applications of deep learning and developing new algorithms. Deep learning is often thought to be possible by simply preparing supervised data, but there are many research areas that require adjustment of many parameters, data shaping, ingenuity in training deep models, and the use of data from other domains.
Examples of research themes
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Research Achievements (Excerpts)
Research Achievements (Excerpts)
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
List of explanations
List of explanations
Deep Learning
Deep Learning
Examples of research themes
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Examples of research themes
Research Achievements (Excerpts)
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Deep Learning
List of explanations
Deep learning generally refers to a deep neural network and is a type of machine learning algorithm. It is used to estimate what is in an image or other specific information, or to predict the future from past time series. Our laboratory conducts joint research with various external institutions and is working on real-world applications of deep learning and developing new algorithms. Deep learning is often thought to be possible by simply preparing supervised data, but there are many research areas that require adjustment of many parameters, data shaping, ingenuity in training deep models, and the use of data from other domains.
Examples of research themes
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Research Achievements (Excerpts)
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
List of explanations
List of explanations
Deep Learning
Deep Learning
Examples of research themes
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Examples of research themes
Research Achievements (Excerpts)
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Deep Learning
List of explanations
Deep learning generally refers to a deep neural network and is a type of machine learning algorithm. It is used to estimate what is in an image or other specific information, or to predict the future from past time series. Our laboratory conducts joint research with various external institutions and is working on real-world applications of deep learning and developing new algorithms. Deep learning is often thought to be possible by simply preparing supervised data, but there are many research areas that require adjustment of many parameters, data shaping, ingenuity in training deep models, and the use of data from other domains.
Examples of research themes
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Research Achievements (Excerpts)
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
List of explanations
List of explanations
Deep Learning
Deep Learning
Examples of research themes
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Examples of research themes
Research Achievements (Excerpts)
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Deep Learning
List of explanations
Deep learning generally refers to a deep neural network and is a type of machine learning algorithm. It is used to estimate what is in an image or other specific information, or to predict the future from past time series. Our laboratory conducts joint research with various external institutions and is working on real-world applications of deep learning and developing new algorithms. Deep learning is often thought to be possible by simply preparing supervised data, but there are many research areas that require adjustment of many parameters, data shaping, ingenuity in training deep models, and the use of data from other domains.
Examples of research themes
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Research Achievements (Excerpts)
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
List of explanations
List of explanations
Automatic recognition of catches
Realizing sustainable fisheries requires appropriate resource surveys, science-based stock assessments, and resource management. However, resource surveys, which form the basis for stock assessments, are often conducted manually by fisheries researchers, and data digitalization is insufficient. This research project is developing a platform to instantly digitize detailed information on catches using image recognition immediately after landing.
The latter cited study addresses the task of classifying fish species down to the species level, not just the family level, using deep learning for fish recognition. It proposes a detailed fish species identification method using multimodal learning with image information and meristic data such as fin ray counts.
Personal identification using dental images
In the event of disasters, many unidentified individuals may occur, necessitating methods for identification. Teeth are resistant to heat and impact, and dental alignment, orthodontic treatment, and dental treatment marks can provide sufficient characteristics to identify individuals. In many cases, authentication is performed visually with the cooperation of dentists, using X-ray and CT scan images. This research project is developing a deep learning method for personal identification using X-ray and CT scan images. In particular, there is a demand for technology development to train models from a small number of training data.
Detection of UHR frequency from plasma wave observation data
In the field of solar-terrestrial science, we are working to elucidate the mechanisms of natural phenomena by analyzing information from space observed by artificial satellites. When analyzing these phenomena, it is necessary to accurately extract specific plasma wave phenomena from a large amount of observational data. However, manually extracting all instances of the target phenomenon from data that is continuously observed 24/7 at high temporal resolution is difficult and can significantly reduce the time available for the actual analysis work. Therefore, we focus on a phenomenon called Upper Hybrid Resonance (UHR) emission, a type of plasma wave phenomenon, and develop a method to automatically estimate it. The observational data consists of time-series changes in frequency spectra, and we use machine learning to solve the regression problem of extracting UHR emissions from the frequency spectrum at each time point.
Deep Learning
Deep Learning
Deep Learning
Examples of research themes
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Examples of research themes
Research Achievements (Excerpts)
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Deep Learning
List of explanations
Deep learning generally refers to a deep neural network and is a type of machine learning algorithm. It is used to estimate what is in an image or other specific information, or to predict the future from past time series. Our laboratory conducts joint research with various external institutions and is working on real-world applications of deep learning and developing new algorithms. Deep learning is often thought to be possible by simply preparing supervised data, but there are many research areas that require adjustment of many parameters, data shaping, ingenuity in training deep models, and the use of data from other domains.
Examples of research themes
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic font generation using GAN, Automatic detection of plasma waves (Joint research: ISAS/JAXA), Fish region detection and fish species estimation in catches (Joint research: Japan Fisheries Research and Education Agency), Research on dental diagnostic image recognition (Joint research: Fukui University Hospital), etc.
Research Achievements (Excerpts)
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Research Achievements (Excerpts)
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
Automatic Electron Density Determination by using a Convolutional Neural Network, IEEE Access, 2019.
"Personal identification with orthopantomography using simple convolutional neural networks: A preliminary study", Scientific Reports, 2020.
Fish Species Identification using a CNN-based Multimodal Learning Method, MLHMI2020, Singapore, 2020.
CNN-based Criteria for Classifying Artists by Illustration Style, MLHMI2020, Singapore, 2020.
List of explanations
List of explanations
Automatic recognition of catches
Realizing sustainable fisheries requires appropriate resource surveys, science-based stock assessments, and resource management. However, resource surveys, which form the basis for stock assessments, are often conducted manually by fisheries researchers, and data digitalization is insufficient. This research project is developing a platform to instantly digitize detailed information on catches using image recognition immediately after landing.
The latter cited study addresses the task of classifying fish species down to the species level, not just the family level, using deep learning for fish recognition. It proposes a detailed fish species identification method using multimodal learning with image information and meristic data such as fin ray counts.
Personal identification using dental images
In the event of disasters, many unidentified individuals may occur, necessitating methods for identification. Teeth are resistant to heat and impact, and dental alignment, orthodontic treatment, and dental treatment marks can provide sufficient characteristics to identify individuals. In many cases, authentication is performed visually with the cooperation of dentists, using X-ray and CT scan images. This research project is developing a deep learning method for personal identification using X-ray and CT scan images. In particular, there is a demand for technology development to train models from a small number of training data.
Detection of UHR frequency from plasma wave observation data
In the field of solar-terrestrial science, we are working to elucidate the mechanisms of natural phenomena by analyzing information from space observed by artificial satellites. When analyzing these phenomena, it is necessary to accurately extract specific plasma wave phenomena from a large amount of observational data. However, manually extracting all instances of the target phenomenon from data that is continuously observed 24/7 at high temporal resolution is difficult and can significantly reduce the time available for the actual analysis work. Therefore, we focus on a phenomenon called Upper Hybrid Resonance (UHR) emission, a type of plasma wave phenomenon, and develop a method to automatically estimate it. The observational data consists of time-series changes in frequency spectra, and we use machine learning to solve the regression problem of extracting UHR emissions from the frequency spectrum at each time point.
Automatic recognition of catches
Realizing sustainable fisheries requires appropriate resource surveys, science-based stock assessments, and resource management. However, resource surveys, which form the basis for stock assessments, are often conducted manually by fisheries researchers, and data digitalization is insufficient. This research project is developing a platform to instantly digitize detailed information on catches using image recognition immediately after landing.
The latter cited study addresses the task of classifying fish species down to the species level, not just the family level, using deep learning for fish recognition. It proposes a detailed fish species identification method using multimodal learning with image information and meristic data such as fin ray counts.
Personal identification using dental images
In the event of disasters, many unidentified individuals may occur, necessitating methods for identification. Teeth are resistant to heat and impact, and dental alignment, orthodontic treatment, and dental treatment marks can provide sufficient characteristics to identify individuals. In many cases, authentication is performed visually with the cooperation of dentists, using X-ray and CT scan images. This research project is developing a deep learning method for personal identification using X-ray and CT scan images. In particular, there is a demand for technology development to train models from a small number of training data.
Detection of UHR frequency from plasma wave observation data
In the field of solar-terrestrial science, we are working to elucidate the mechanisms of natural phenomena by analyzing information from space observed by artificial satellites. When analyzing these phenomena, it is necessary to accurately extract specific plasma wave phenomena from a large amount of observational data. However, manually extracting all instances of the target phenomenon from data that is continuously observed 24/7 at high temporal resolution is difficult and can significantly reduce the time available for the actual analysis work. Therefore, we focus on a phenomenon called Upper Hybrid Resonance (UHR) emission, a type of plasma wave phenomenon, and develop a method to automatically estimate it. The observational data consists of time-series changes in frequency spectra, and we use machine learning to solve the regression problem of extracting UHR emissions from the frequency spectrum at each time point.