Development of Pneumonia Severity Scoring Technology with Thoracic Radiation Imaging
Jul 16, 2024
Inha University announced on the 16th that a medical artificial intelligence lab led by Professor Lee Hyun-kyu of the Department of Medicine has developed a technology that can score the severity of pneumonia on chest radiographic images.
Lee Jong-beop, a master's student in the Department of Electrical and Computer Engineering under the research team, will present a thesis on the subject of the 'COVID19 to Pneumonia: Multi-Region Lung Severity Classification Network' (COVID19 to Pneumonia: Multi-Region Lung Severity Classification Network) using the CNN Transformer Position-Aware Feature Encoding Network as the first author.
This paper proposed a network that incorporates each area information of the lungs and applies location recognition to distinguish features with similar patterns in different lung areas but different causes.
Existing studies have only briefly provided the presence or absence of pneumonia in chest radiographic images, while segmenting the severity of pneumonia in this study is an important achievement. In particular, it performs well on large COVID-19 datasets as well as ICU datasets at Inha University Hospital, demonstrating its effective performance on small or imbalanced datasets.
The research team will present these achievements at the Medical Image Computing and Computer Assisted Intervention 2024, the best academic conference in the field of medical imaging and medical imaging computing in Morocco in October.
Co-author Kim Jeong-soo, professor of respiratory medicine at Inha University Hospital, explained, "Unlike previous studies, this study has a special advantage in that it allows quantitative indication and follow-up of changes in lung conditions of pneumonia patients admitted to intensive care units.".
Lee Hyun-kyu, a professor of medicine at Inha University, said, `The medical artificial intelligence laboratory has continued to study pragmatism centered on real users through joint research with clinicians""We plan to continue to focus on commercialization beyond research using this achievement recognized by the academic community as an opportunity. "
Lee Jong-beop, a master's student in the Department of Electrical and Computer Engineering under the research team, will present a thesis on the subject of the 'COVID19 to Pneumonia: Multi-Region Lung Severity Classification Network' (COVID19 to Pneumonia: Multi-Region Lung Severity Classification Network) using the CNN Transformer Position-Aware Feature Encoding Network as the first author.
This paper proposed a network that incorporates each area information of the lungs and applies location recognition to distinguish features with similar patterns in different lung areas but different causes.
Existing studies have only briefly provided the presence or absence of pneumonia in chest radiographic images, while segmenting the severity of pneumonia in this study is an important achievement. In particular, it performs well on large COVID-19 datasets as well as ICU datasets at Inha University Hospital, demonstrating its effective performance on small or imbalanced datasets.
The research team will present these achievements at the Medical Image Computing and Computer Assisted Intervention 2024, the best academic conference in the field of medical imaging and medical imaging computing in Morocco in October.
Co-author Kim Jeong-soo, professor of respiratory medicine at Inha University Hospital, explained, "Unlike previous studies, this study has a special advantage in that it allows quantitative indication and follow-up of changes in lung conditions of pneumonia patients admitted to intensive care units.".
Lee Hyun-kyu, a professor of medicine at Inha University, said, `The medical artificial intelligence laboratory has continued to study pragmatism centered on real users through joint research with clinicians""We plan to continue to focus on commercialization beyond research using this achievement recognized by the academic community as an opportunity. "
|
bellho@sportschosun.com
