Overview of Deep learning

The paper looks at both the history (quickly) and gives an overview of “Deep Learning”.

In their words:

The History: Birth, Decline and Prosperity

I skimmed most of the overview because it was not what I came to the paper for.

I think this is more interesting…

Outline

The paper doesn’t explicitly call this the outline, but for what I’m interested in, its the general progression. The paper reviews two main object detection methods: Region proposal and regression/classification. I want to spend a couple days digging into the specifics of these different approaches. There are 4.5 pages dedicated to the region frameworks.

Region proposal based frameworks R-CNN, SPP-net, Fast R-CNN, Faster R-CNN, R-FCN, FPN, and Mask R-CNN.

Regression/Classification Frameworks Multibox, AttentionNet, G-CNN, YOLO, SSD, YOLOv2, DSSD, and DSOD.