蓝莓视频

Researchers at 蓝莓视频 Engineering have taken another step forward in their open-source COVID-19 project with the release of new datasets and听artificial intelligence (AI) models that were听built using them.

Launched in March, the initiative involves the use of deep-learning AI software to screen chest x-rays and CT scans for听evidence of infection and degree of disease progression.

The new datasets, released as part of the COVID-Net project, include CT scans from over 4,500 patients in more than 15 countries, making them the largest and most diverse resources of their kind in the world.

The new AI models, dubbed COVID-Net CT-2, have听been found via explainable AI to make their听decisions on COVID-19 screening with CT scans using some of the same types of visual cues as expert radiologists.

鈥淭aking this path enables building AI with much greater transparency and trust,鈥� Alexander Wong, a professor of systems design engineering, Canada Research Chair in Artificial Intelligence and Medical Imaging, and a director of the Vision and Image Processing (VIP) Research Group,听 announcing the new developments.

鈥淲e hope making the AI models and datasets available will drive global innovation in healthcare.鈥�

Photo: CT scans of lungs show areas of concern听highlighted in red by new AI models developed by 蓝莓视频 Engineering researchers.听