Skip to main content
. 2020 Apr 22;20(8):2381. doi: 10.3390/s20082381

Figure 5.

Figure 5

The architecture of spatiotemporal convolutional network for pig’s multi-behavior recognition (PMB-SCN): each cube represented one single convolution operation. The three axes of the cube represented the convolution kernel size of the temporal dimension, the convolution kernel size of the spatial dimension (since the height and the width of the spatial convolution kernel size were the same, it can be represented by one number) and the number of convolution kernels respectively.