High-Resolution 3D Reconstruction Using Photoacoustic and Ultrasound Data
Dramatic Advances in Diagnostic Accuracy and On-Site Usability
A research team at Pusan National University (President Choi Jaewon) has developed a technology that uses artificial intelligence to generate three-dimensional images at MRI-level quality from photoacoustic and ultrasound imaging data, drawing significant attention.
This technology enables high-resolution, real-time visualization of complex human anatomical structures and is being evaluated as a next-generation diagnostic tool that can overcome the limitations of conventional 2D ultrasound imaging.
The team, led by Professor Minwoo Kim of the Department of Biomedical Convergence Engineering at Pusan National University, announced on June 24 that they have implemented a technology that precisely reconstructs anatomical 3D structures by automatically tracking the movement of the probe using an AI-based "MoGLo-Net (Motion-based Global-Local Network)" model.
Professor Minwoo Kim, PhD candidate Siyeol Lee. Courtesy of Pusan National University
Conventional ultrasound equipment provides images in a 2D (planar) format, requiring physicians to manually move the device to scan the desired area. However, because the human body is three-dimensional, there have been limitations in achieving accurate diagnoses. In particular, when using a freehand method to move the ultrasound probe, a separate position sensor is required for precise 3D reconstruction, which leads to increased equipment size, reduced accuracy, and user inconvenience.
To address these issues, the Pusan National University research team introduced a technology that enables AI to precisely analyze subtle movements between ultrasound and photoacoustic images, automatically infer the probe’s position, and reconstruct 3D images in real time?all without external sensors. This technology is highly applicable in clinical settings because it allows for the creation of precise three-dimensional structures using only compact ultrasound devices, without the need for expensive and bulky MRI equipment.
The research team also demonstrated clinical feasibility by presenting 3D examples that clearly visualize even microvessels, using "photoacoustic imaging" technology that combines ultrasound and lasers.
Professor Minwoo Kim explained, "Conventional ultrasound results can vary depending on the user's level of experience, but this new technology enables automated, precise 3D imaging, which reduces reliance on the operator's expertise and improves both consistency and accuracy in diagnosis."
He added, "This is the world’s first attempt to implement freehand photoacoustic imaging in 3D, and the technology has very high potential for commercialization as a next-generation autonomous ultrasound diagnostic system."
This technology can be easily used in various medical environments such as diagnostic rooms, emergency sites, and operating rooms. It is particularly valued as a practical alternative or supplement to expensive MRI equipment, due to its high image precision and analytical capabilities, as well as its relatively low cost and excellent portability.
This research was conducted with PhD candidate Siyeol Lee of the Department of Information Convergence Engineering and AI as the first author, alongside Professor Kim, and was published in the international journal IEEE Transactions on Medical Imaging.
This study was supported by the Mid-Career Researcher Program of the Ministry of Science and ICT and the National Research Foundation of Korea, as well as the AI Convergence Innovation Talent Development Program and the Generative AI Leading Talent Development Program of the Institute of Information & Communications Technology Planning & Evaluation.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
