Skip to content

Latest commit

 

History

History
57 lines (37 loc) · 1.57 KB

ProjectionPoolingLayer.md

File metadata and controls

57 lines (37 loc) · 1.57 KB
Raw

CProjectionPoolingLayer Class

This class implements a layer which performs mean pooling over one of the blob dimensions.

Settings

Projection dimension

// Projection dimension
void SetDimenion( TBlobDim dimension );

The default value is BD_Width.

Restore original image size

void SetRestoreOriginalImageSize( bool flag );

If true then output blob will be of the same size as input blob, and mean values will be broadcasted along pooling dimension. If false then output blob size along pooling dimension will be 1. The default value is false.

Trainable parameters

The layer has no trainable parameters.

Inputs

The single input accepts a blob of the following dimensions:

  • BatchLength * BatchWidth * ListSize - the number of images in the set
  • Height - images' height
  • Width - images' width
  • Depth * Channels - the number of channels the image format uses

Outputs

The single output contains a blob with the results.

If GetRestoreOriginalImageSize is true then output is of the same size as input.

If GetRestoreOriginalImageSize is false then the projection dimension of output size is equal to 1 and the rest of the dimensions are equal to the ones of the input.