This repository contains pre-trained models for the algorithms contained in this repository. See that repository for further information, including a section on how to use these models. These are intended to be used as a demonstration of the algorithm.
This repository contains the following files:
structurelist- Contains a list of structures used when training the models.filter_models- Contains particle filter definition files that define the prediction potentials.forest_models- Contains random forest models that define the observation potentials. There are several files needed to define a single model, with each performing a different task (classification or cardiac phase regression) and/or operating on different image orientations. There are two such models here, one using rotation invariant features (therotinvdirectory) and another using rectangular features (therecdirectory).