A research team from Vietnam National University Ho Chi Minh City has received its first European patent for an AI technology designed to improve medical image annotation.
The European Patent Office (EPO) has granted the patent "Annotation of Slice Images in Volumetric Medical Images" to a research team from the University of Science, Vietnam National University Ho Chi Minh City (VNU-HCM).
European patent for medical AI
The patent was published in the European Patent Bulletin on July 8, 2026, marking a new milestone in research, innovation and international intellectual property for the University of Science as well as Vietnam National University Ho Chi Minh City.
It is VNU-HCM's first European patent in the field of artificial intelligence (AI) applied to medicine.
The invention was led by Dr. Le Khanh Duy of the Faculty of Information Technology, together with Associate Professor Tran Minh Triet, Vice Rector of the University of Science, and two alumni from the university's 2018 Honors Bachelor's Program, Nguyen Ho Thang Long and Pham Minh Khoi, who are currently doctoral researchers at Dublin City University in Ireland.
The four Vietnamese scientists awarded the European patent.
According to Dr. Le Khanh Duy, the research addresses one of the fundamental challenges in building AI systems for healthcare: annotating CT and MRI images to create high-quality training datasets.
Current medical imaging datasets often contain hundreds or even thousands of image slices. Specialists must accurately identify organs or lesions in every individual image, a process that is both time-consuming and prone to inconsistencies across adjacent slices, ultimately affecting dataset quality.
To overcome these limitations, the research team developed a solution that assists with annotating three-dimensional medical images by combining computer vision algorithms with advanced data visualization techniques. Once a user labels or corrects a single image slice, the system automatically predicts and propagates annotations to neighboring slices while displaying a confidence score for each prediction, allowing users to verify and adjust the results when necessary.
This approach significantly reduces annotation time while improving the consistency of training datasets, a key factor in the performance of AI models used for medical image analysis and diagnosis.
Dr. Le Khanh Duy said the core value of the invention lies not in creating a new AI algorithm but in proposing a more effective interaction model between human experts and AI-assisted annotation systems.
"The solution enables users to process large volumes of medical images more efficiently, thereby improving the creation of datasets for developing artificial intelligence models," he said.
The research team hopes the technology will be integrated into medical image annotation software, contributing to the development of more high-quality datasets for research and next-generation AI systems used in medical imaging and diagnostics.
Two international patents in two months
Four Vietnamese scientists receive a U.S. patent.
Earlier, four Vietnamese scientists were also granted a patent by the United States Patent and Trademark Office (USPTO) for a technology that fabricates memristors using chromium oxide (CrOₓ), opening new possibilities for next-generation AI hardware.
The inventors include Associate Professor Pham Kim Ngoc and Master Pham Phu Quan from the University of Science, Professor Phan Bach Thang from the University of Health Sciences (VNU-HCM), and Associate Professor Nguyen Tran Thuat from the Institute of Semiconductors and Advanced Materials at Vietnam National University Hanoi.
The patent, titled "Memristor Structures with Analog Switching Characteristics and Method for Fabricating the Same," was granted by the USPTO on June 30, 2026.
A key innovation of the research is the development of a chromium oxide (CrOₓ)-based materials platform instead of commonly used materials such as TiO₂, HfOₓ or TaOₓ. The platform can be flexibly combined with various oxide layers and metal electrodes to create new generations of memristors.
According to the research team, the technology enables the development of multistate memory devices with nonlinear switching characteristics, providing the foundation for artificial synapses used in brain-inspired computing systems.
The patent also proposes an integrated memristor array architecture in which each device both stores and processes data locally, eliminating the need for continuous data transfer between memory and processors. This approach paves the way for In-Memory Computing and Neuromorphic Computing chips, enabling faster and more energy-efficient AI processing.
The patent is the result of many years of research and further demonstrates the growing capabilities of Vietnamese scientists in semiconductor research on the international stage.
Within a short period, Vietnamese researchers - most of them affiliated with Vietnam National University Ho Chi Minh City - have secured two major international patents, one from the United States and another from Europe.
Vietnamnet.vn
Podcast
.svg){kind=icon}
