Skip to main content
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub
The LatentOperationSharpen node applies a sharpening effect to latent representations using a Gaussian kernel. It works by normalizing the latent data, applying a convolution with a custom sharpening kernel, and then restoring the original luminance. This enhances the details and edges in the latent space representation.

Inputs

ParameterData TypeRequiredRangeDescription
sharpen_radiusINTNo1-31The radius of the sharpening kernel (default: 9)
sigmaFLOATNo0.1-10.0The standard deviation for the Gaussian kernel (default: 1.0)
alphaFLOATNo0.0-5.0The sharpening intensity factor (default: 0.1)

Outputs

Output NameData TypeDescription
operationLATENT_OPERATIONReturns a sharpening operation that can be applied to latent data