Skip to main content
. Author manuscript; available in PMC: 2023 Feb 24.
Published in final edited form as: IEEE Trans Big Data. 2016 May 30;6(2):322–333. doi: 10.1109/tbdata.2016.2573280

Fig. 3.

Fig. 3.

Detailed architecture of the VGG model. “Convolution”, “Max pooling” and “ReLU” denote convolutional layer, max pooling layer and rectified linear unit function layer, respectively. This model consists of 36 layers. We extracted features from layers 71, 21, 24, and 30.