SpatialDropout

Introduced by Tompson et al. in Efficient Object Localization Using Convolutional Networks

SpatialDropout is a type of dropout for convolutional networks. For a given convolution feature tensor of size $n_{\text{feats}}$×height×width, we perform only $n_{\text{feats}}$ 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.

Source: Efficient Object Localization Using Convolutional Networks

Read Paper See Code

Papers

Paper	Code	Results	Date	Stars

Tasks

Task	Papers	Share
Semantic Segmentation	13	24.53%
Autonomous Driving	5	9.43%
Quantization	2	3.77%
Autonomous Vehicles	2	3.77%
Real-Time Semantic Segmentation	2	3.77%
Image Classification	2	3.77%
Image Segmentation	2	3.77%
Instance Segmentation	2	3.77%
Scene Understanding	2	3.77%

Usage Over Time

This feature is experimental; we are continuously improving our matching algorithm.

Components

Component	Type	Add Remove
🤖 No Components Found	You can add them if they exist; e.g. Mask R-CNN uses RoIAlign

Categories

Add Remove

Regularization