In MeVisLab, a variety of image processing concepts is available. They differ in scope:
Random Access (Tile requesting)
Sequential Image Processing
All those concepts are discussed in detail in the ML Guide, chapter “Image Processing Concepts”.
When choosing your approach, keep in mind that some of the concepts are not scaling well for larger images. For example, the page-based approach can only be beneficial if the pages are of a size so that they actually fit into memory, or can be administered by the internal ML host / cache. Always try to set the page sizes to some reasonable values, like 128x128x1x1x1x1. You can do this with
ImagePropertyConvert modules (insert them right after the loading modules in your network).
The ITK modules frequently produce memory allocation problems for large images because they try to load the entire image at once. You can find out about the memory management in the ITK module help. Look for something like
© 2021 MeVis Medical Solutions AG