Alibaba has opened the source code of the Damo Radar medical model

Company Alibaba has published the source code of the neural network Damo Radar, designed for automated analysis of medical images. The model is capable of detecting nearly 150 different pathologies, including cancer, based on computed tomography (CT) data.

According to the results of a study published in the journal Science, the system demonstrates a recognition accuracy of 91.3%. In comparative tests, the algorithm showed higher efficiency in diagnosis than professional radiologists participating in the experiment.

The development is aimed at supporting clinical decisions and accelerating primary patient screening. The use of the model allows reducing the time for processing images and decreasing the likelihood of diagnostic errors associated with the human factor in analyzing large amounts of data.

Opening access to the Damo Radar architecture allows medical organizations to integrate the solution into their own image processing systems. The technology is intended to become a decision support tool, automating routine stages of CT scan analysis in conditions of high workload on medical personnel.