Researchers from the Air Force Medical University in China, Tsinghua University and SenseTime, which develops artificial intelligence technologies, found that cancer pathological images has developed a versatile artificial intelligence model that can read in detail. In this study, approximately 300,000 digital pathology images were used, compiled from local data in the country.

As a result of the research, the artificial intelligence model was created more than twenty organs It was said that he could analyze it. It was stated that these analyses included items such as pancreas classification, lesion definition, multiple cancers, and biomarker evaluation.

The success rate is 95 percent

researchers PathOrchestra The versatile artificial intelligence model, named after the application, has been used in nearly 50 cancer cases. more than 95 percent He also managed to achieve high accuracy. Those who spoke about the research said that this research is an indicator of China’s rapid growth in the field of artificial intelligence.

The artificial intelligence model trained itself using a method called cross-learning. From the Air Force Medical University Academy of Basic Medical Sciences Professor Wang Zhe, He stated that detecting and interpreting pathological images is very difficult due to the differences between pathogens, and therefore examining diseased structures is considered the most difficult problem in the field of image processing.

As this model is further developed The Burden on Pathologists It is expected that this will be relaxed considerably.