(Health Korea News / Yu Ji-in) Seoul National University Hospital Consortium announced on the 26th that it is carrying out a project to build ‘pediatric abdomen multimodal and synthetic data’ to expand ultra-large-scale AI and respond to on-site demands.
Seoul National University Hospital expects that this data construction project will improve the diagnostic accuracy of pediatric abdominal diseases, thereby improving the health management and treatment outcomes of pediatric patients.
This project is part of the ‘2024 Ultra-Large AI Diffusion Ecosystem Creation Project’ hosted by the Ministry of Science and ICT and promoted by the National Intelligence Information Society Agency (NIA). The Seoul National University Hospital Consortium was selected as the final implementing agency for the ‘Pediatric Abdominal Multimodal and Synthetic Data’ construction project in the healthcare field. The data construction project will be carried out until the end of this year, and after the project is completed, it will be released through the ‘AI-Hub’ operated by the NIA.
This consortium, hosted by Seoul National University Hospital, is comprised of seven institutions: ▲Kyungpook National University Industry-Academic Cooperation Foundation ▲Korea University Industry-Academic Cooperation Foundation ▲Gil Medical Foundation ▲Yangsan Busan National University Hospital ▲Urban Data Lab ▲Sur, and its budget is 1.2 billion won.
A commencement report meeting to announce the full-scale start of the project was held on the 11th at Seoul National University Children’s Hospital together with NIA. At the event, the Seoul National University Hospital Consortium emphasized the importance of enhancing AI service competitiveness and the need to build high-quality, large-scale data at the national level.
The data constructed in this project consists of X-ray and medical image data for diagnosing pediatric abdominal diseases. The data is divided into two types: multimodal data and synthetic data.
Multimodal data includes approximately 2,000 pairs of X-rays and other medical images and clinical data from the same patient. Synthetic data is based on real multimodal data and is expected to be constructed with approximately 10,000 data that have been labeled with major clinical symptoms, diagnoses, and treatment methods. These data will be used to develop an artificial intelligence model that assists in the diagnosis of pediatric abdominal diseases.
The Seoul National University Hospital consortium said, “We plan to collect multimodal data, combine various information, and create a more accurate model,” and “Based on this, we will be able to increase data diversity and solve the problem of data shortage by creating synthetic data through labeling work and artificial intelligence techniques.”
Professor Hyunyoung Kim (Pediatric Surgery) of Seoul National University Hospital, who is in charge of this project, said, “The super-large AI data construction project will be a new leap forward in the diagnosis and treatment of pediatric abdominal diseases by utilizing AI technology,” and “I expect that the accuracy and efficiency of pediatric abdominal disease diagnosis will be greatly improved through this project.”
Copyright © Health Korea News. Unauthorized reproduction and redistribution prohibited.
Source: www.hkn24.com
