Viet-Anh on Software Logo

What is: SpatialDropout?

SourceEfficient Object Localization Using Convolutional Networks
Year2000
Data SourceCC BY-SA - https://paperswithcode.com

SpatialDropout is a type of dropout for convolutional networks. For a given convolution feature tensor of size n_featsn\_{\text{feats}}×height×width, we perform only n_featsn\_{\text{feats}} dropout trials and extend the dropout value across the entire feature map. Therefore, adjacent pixels in the dropped-out feature map are either all 0 (dropped-out) or all active as illustrated in the figure to the right.