Skip to main content
Advertisement
Partitioning variability in animal behavioral videos using semi-supervised variational autoencoders
Table 3
Convolutional VAE architecture for the IBL dataset using N latents.
Kernel size and stride size are defined as (x pixels, y pixels); padding size is defined as (left, right, top, bottom); output size is defined as (x pixels, y pixels, channels). Reparameterization details not included.