AdaptiveKernelType

The 'Box' kernel treats all pixels in the window equally when calculating the local average.

• Fastest option, suitable for most standard documents.
• May be less effective at suppressing noise in highly variable backgrounds.

Example:
Use for forms, invoices, or documents with generally uniform backgrounds.

The 'Gaussian' kernel gives more weight to pixels near the center of the window, producing smoother transitions and reducing the impact of outliers.

• Produces higher quality results for documents with gradients, shadows, or background patterns.
• May be slower than 'Box', but can reduce noise and improve text clarity in challenging images.

Example:
Use for photographs, receipts, or documents with significant background variation.