Researchers at the Chinese University of Hong Kong (CUHK) have developed an artificial intelligence (AI) model that could help address clinical ophthalmology tasks such as disease detection and diagnosis.
The model, called VisionFM, has been shown to be effective in diagnosing and predicting various eye pathologies, which could spur the implementation of new clinical applications based on complementary data, according to a report, published on Saturday by local newspaper South China Morning Post, about a study published last month in the journal NEJM AI.
The study showed that the technology performs comparable to or better than mid-level ophthalmologists in diagnosing 12 eye diseases.
Furthermore, it outperformed the foundational model in this domain, RETFound, in predicting glaucoma progression, the researchers indicated.
Unlike many AI systems that focus on specific pathologies, VisionFM is proposed as an open-source generalist model, capable of adapting to additional datasets and new imaging modalities.
The CUHK team of scientists said the VisionFM architecture was trained using 3.4 million images from half a million patients.
One of the limitations of this system lies in the higher accuracy rate when analyzing data from China, since the training data set from this region is larger than those from other areas, he highlighted.
The study also focused on comparing the performance of this tool with ophthalmologists from the Asian giant, which suggests the need for more extensive and diverse future research involving experts from different countries.
Leave a Reply