Google has won a contract with the Defense Innovation Unit (DIU) of the US Department of Defense. It will use recent technological advances to create a Google microscope equipped with AI and AR that will improve the accuracy of detecting cancer.
About 5% of cancer diagnoses are made incorrectly, and in half of these cases, the failure to detect the cancer causes subsequent problems.
Google plans to train an artificial intelligence using the open source platform TensorFlow to detect cancer cells in photos taken by a microscope. The Mountain View company has already started developing neural networks using the Google Cloud Healthcare API to anonymize and segment existing datasets. When the AI is fully trained, Google will design its AI microscope with a built-in AR (augmented reality) overlay to show doctors information about how likely cells are to be cancerous.
A magic eye for cancer detection
“To detect cancer and treat it effectively, speed and accuracy are key.” He states it Mike Daniels, vice president of Google Cloud for the global public sector.
We partner with DIU to deliver our machine learning and artificial intelligence technology. We will help frontline health workers improve the lives of our military men and women and their families.
Mike Daniels, vice president of Google Cloud for the global public sector.
Google's AI and AR microscope: the roadmap
Google hopes its method will reduce the huge volume of data doctors have to deal with and make diagnoses faster and cheaper, as well as more accurate.
The first phase will see the selection of Defense Health Agency treatment facilities for research purposes only. In the second stage, the technology will be rolled out to U.S. military health system hospitals for actual real-world testing.