Published 20 Nov.2025 10:16(KST)
Deepnoid, a first-generation domestic medical artificial intelligence (AI) specialist company, announced on November 20 that five of its research abstracts have been accepted for presentation at 'RSNA 2025'.
The annual academic conference, organized by the Radiological Society of North America, will be held at McCormick Place in Chicago, USA, for five days from November 30 to December 4, local time.
The Radiological Society of North America, established in 1915, is the world's largest radiology society. It has more than 51,000 members from 157 countries. Last year, over 700 companies and 38,000 registered participants from around the world took part in the event.
A total of five abstracts were accepted, each addressing different modalities. However, all share the common objective of advancing the reliability, interpretability, and clinical reproducibility of medical AI models. These studies have been recognized for their academic significance, as they go beyond simply improving AI model performance and demonstrate the potential for solving real clinical problems and practical use.
The individual abstracts cover: a dual-learning architecture to address performance gaps between PA and AP chest X-ray imaging; an AI-assisted double reading study using a vision-language model (VLM) to detect chest X-ray interpretation errors; an evaluation of visualization performance and clinical utility of chest X-ray reports using heatmaps generated by a vision-language model; a study comparing the performance of an aneurysm auto-detection and segmentation model developed by additional training of a general-purpose segmentation AI model (SAM); and a multicenter clinical trial evaluating the diagnostic performance of the DEEP:LUNG pulmonary nodule CADx system.
On December 1 (local time), Deepnoid will also give a presentation at the 'RSNA 2025 Seminar', organized by the Korea Health Industry Development Institute as part of the conference program. Choi Kyo-yoon, Senior Researcher at Deepnoid AI Research Institute's Advanced AI Technology Team, will deliver a lecture on 'Next-Generation Modalities'.
Deepnoid will operate an exhibition booth at RSNA from November 30 to December 3. The booth will feature demonstrations for global medical professionals, hospital representatives, and AI researchers. Through booth operations, the company aims to share academic achievements and lay the foundation for entry into the global market.
Choi Woosik, CEO of Deepnoid, said, "The acceptance of multiple abstracts means that Deepnoid's technological capabilities have been internationally recognized at RSNA, a global stage leading the medical AI trend. Through the conference, we will directly communicate with potential global customers and gather feedback, while also actively seeking collaboration opportunities with overseas research institutions."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.