History

Note: in the video, it should be "Logistic and statistical pattern recognition"

Summary

Logistic and statistical pattern recognition were first applied in medicine in the 1960s, but not for medical imaging. In 1980s, computers became more prevalent and they changed how the doctor would approach medical imaging problems. MRI was first applied clinically in the 1980s as well. Many classic machine learning models like support vector machines and random forests were not even invented until the mid-to-late 90s, but they quickly made their way into healthcare, with some specific use cases for medical imaging. 2012 saw the real promise for deep learning in general imaging problems, however, with the win of deep learning on the ImageNet challenge from there, it only took 5 years for the first deep learning algorithm to be FDA cleared for medical imaging. Now, in 2020, we bring you the first Udacity course in AI for 2D imaging!

For more information about the history of AI for medical imaging, read this article and this article