Operator Library: Base
The operator SelectSubKernel extracts a rectangular subset of the kernel matrix at the input link. Similar to the ROI an offset for the first element of the new kernel can be defined using parameters FirstROW and FirstCOL. The size (width/height) of the new kernel can be chosen at the output link by setting the kernel size of the link.
Often, this operator is used to extract the central element of a kernel which mostly is the original pixel value.
FirstROW | |
---|---|
Type | static parameter |
Default | 0 |
Range | [0, InputKernelRows - OutputKernelRows] |
This parameter defines the first row for the new kernel. |
FirstCOL | |
---|---|
Type | static parameter |
Default | 0 |
Range | [0, InputKernelColumns - OutputKernelColumns] |
This parameter defines the first column for the new kernel. |
The use of operator SelectSubKernel is shown in the following examples:
-
Synchronization - Avoiding deadlocks.
-
Examples - A convolution with high intensity spot coefficients is made. For results above threshold, the respective pixels are dyed in red.
-
'Nearest Neighbor Demosaicing'
Examples - Nearest Neighbor Bayer Demosaicing
-
'Edge Sensitive Bayer Demosaicing Algorithm'
Examples - Edge Sensitive Laplace Bayer Demosaicing filter
-
'Bayer Demosaicing Algorithm According to Laroche'
Examples - Laroche Bayer Demosaicing filter
-
'Modified Laroche Bayer Demosaicing Algorithm '
Examples - Ressource Optimized Laroche Bayer Demosaicing filter
-
Examples - A binary eroded image is compared with the original. An edge is detected if both differ.
-
Examples - Shows how to extract a sub kernel from a filter to obtain the original image data. This example performs a simple local adaptive binarization.
-
'High Boost Sharpening Filter'
Examples - A high boost Laplace filter for sharpening
-
Examples - The examples shows an automatic dead pixel detection and replacement.