Densely Connected Convolutional Networks

It has been shown that convolutional networks can be substantially deeper, more accurate, and efficient to train if they contain shorter connections between layers close to the input and those close to the output. The authors of this paper take this insight to its logical conclusion by introducing the Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward fashion which comes with advantages such as alleviating the vanishing-gradient problem, strengthening feature propagation, encouraging feature reuse, and substantially reducing the number of parameters.


Want to receive more content like this in your inbox?