Fig. 1
From: Deep learning for DNase I hypersensitive sites identification
The internal structure of the SPP layer: Number of 384 describes the number of the fifth convolution layer. The fifth layer convolution is the last layer. The output obtained from the fifth layer convolution layer are pooled through the pooling layer of n/4, n/2, and n/1 dimension respectively, then the 4, 2, and 1 features are obtained. Finally, the 4+2+1=7 features are gathered into the full connection layer