Skip to main content
. Author manuscript; available in PMC: 2017 Jan 1.
Published in final edited form as: Neuroimage. 2015 May 15;124(0 0):127–146. doi: 10.1016/j.neuroimage.2015.05.018

Figure 3.

Figure 3

Feature vector representation from the trained DNN weights. The learned feature vector, F(k)(J+1), at the kth hidden node in the (J+1)th hidden layer is defined from a linear combination of the feature vectors in the Jth hidden layer (please refer to the detailed description in the Methods section). Input layer is defined as the 0th layer.