Unfortunately this site is still a work-in-progress.
Here’s an interesting video about both medical imaging and deep learning:
Deep Learning in Medical Imaging – Ben Glocker
A good way to keep track of the current exponentially growing interest in machine learning topics is via Google Trends. The following graphs are live, updating with the most recent data. The first one shows the growing interest in machine learning. The CUDA language, developed by Nvidia for high-performance computation on GPU, had enormous interest in the past but has plateaued. This is likely due to the prevalence of frameworks like Tensorflow that now allow developers to ignore the complexity of dealing with CUDA directly.
It’s always interesting to see if one of the Deep Learning frameworks is “winning” the war of hearts and minds of developers. Here I compared some company sponsored frameworks, namely Tensorflow (Google), Core ML (Apple) and CNTK (Microsoft), with more research oriented frameworks e.g. PyTorch, Keras. I’m specifically looking at the time period since the big bang when the current leader, Tensorflow, was introduced in late 2015. It appears that Tensorflow is still the leader. It will be interesting to see if the announcements this summer eventually translate into more support for the Microsoft/CNTK framework, but Google’s Tensorflow has an enviable lead and high-level frameworks like Keras support it. By the way, Apple seems to be a no-show in the DL framework race with Core ML trending at a nearly invisibly low level. Facebook’s Torch is probably some multiple of PyTorch, but it is exceedingly difficult to search for it specifically without dragging in unrelated articles.
And here’s a fun video that has absolutely nothing to do with deep learning or medical imaging: