| Customize Help

Preprocessing the Pattern Matching context


When you are ready to search for the allocated model (either manually or automatically), you must preprocess the Pattern Matching context. The preprocessing stage uses the known model(s) to decide on the optimal search strategy for subsequent search operations. Preprocessing should be performed after all search constraints have been set. Use the MpatPreprocess() function to preprocess the Pattern Matching context.

MpatPreprocess() has a parameter that allows you to specify a typical target image. Providing a typical image is optional; you can set this parameter to M_NULL. If you provide this image, it helps MpatPreprocess() improve the search's robustness and optimize the strategy for subsequent search operations. You should only specify a typical image if the model occurrence will always appear on the same type of background.

Note that when you load or restore a Pattern Matching context (using MpatRestore()), you must preprocess it, even if it was preprocessed before saving.