Custom 3x3 Kernel


The Custom 3x3 Kernel command convolves a user-specified 3x3 kernel. This is a simplified version of the Custom Filter Kernel. The filter kernel contains numeric weights applied to values in the neighborhood of each target pixel. The kernel values are actually weights applied to pixels in a 3x3 neighborhood centered on each target pixel in the image. The kernel weights may be normalized to a "volume" of 1.0 without typing in normalized values.

Properties of the Custom 3x3 Kernel

Profile

Selects the parameter profile for this command.

[Select]

Selects the source images from an image window, folder, files, pattern, or a file list.

Filter Weights

The weight for each element of the filter kernel. After entering weights, you can have the filter normalized during processing checking the Normalize check box.

 Normalize

Check this box to have the filter normalized to unit volume such that the sum of all weights equals 1.0. This is done at processing time so the filter can be shown more simply in the weight grid.

Source Region

Specifies the rectangular region to be processed. The options are as follows:

Entire image: The entire image is processed.

Image Cursor: The region to be processed is selected using the Image Cursor .

Rectangle: The region specified by Min and Max limits for Column and Row coordinates is processed.

Column and Row, Min and Max: When Rectangle is selected, these values specify the column and row limits to be processed.

Tip

Be careful not to use Normalize when the filter weights sum to 0. Many operators such as gradient filters normalize to 0 whereas a sharpening filter, for example, normalizes to a value of 1.

Related Topics

Spatial Filter Commands, Custom Filter Kernel