Beginning Anomaly Detection Using

 Spatial Dropout 2D

keras.layers.SpatialDropout2D()

This function is similar to the spatial dropout 1D function, except it works on 2D 

feature maps. Images can have three channels if they’re color images. They can be 

RGB (red, green, blue), BGR (blue, green, red), HSV (hue, saturation, value), etc., so 

the dimensions of these images are actually (height, width, channels) if it’s formatted 

channels last or (channels, height, width) if it’s formatted channels first.

This function takes one additional parameter compared to SpatialDropout1D():

•