Michelle Horton wrote an interesting article as part of the NVIDIA Developer program that caught my attention as it relates to breast cancer. Cancer is such a heinous disease – and one that could actually be impacted by the use of Artificial Intelligence.
I’m interested in what the computer did to help, because it’s applies to A.I. in general and where and how it’s useful to society. I am not a radiologist, but I know some cancerous masses take certain shapes. When scanned properly a radiologist can tell if a mass is a cancerous or not depending on the physical attributes.
In reality, this in itself isn’t that difficult. In fact NVIDA’s Jetson TX2 comes pre-loaded with image detection software. I haven’t dug into all of it but basically the idea on this stuff is that you tell the machine what to look for and then it goes and looks for it.
When it detects a shape (or pattern within a shape) and can cross reference that with what it knows and make a decision based on that match.
Why would this be exciting for cancer detection? In the future you could self-assess yourself. In the case of this study, the researchers found that it could reduce the need for unnecessary biopsies, which in itself is a huge step forward.
To learn more about this specific research, which involved more than simple image detection you can read the study as published on Science.org.