Update: I wrote a new EvalFun which saves the femurs from the Crossvalidation.leaveOneOutCrossvalidation function so I can run my own evaluations.
However, as a test I computed the Dice score between those femurs and their corresponding original femurs, and the results were different from when I directly included the Dice score as a part of the EvalFun.
I'm not sure what might be causing this so far. Is this something that might be expected?
Any help is appreciated.
Thanks,
Jasper