I am using cv::ml::Boost for 2d range data person detection and when I try to set the parameter cv::ml::DTrees::setCVFolds(int value) bigger than 1, there is always occuring a segfault.
So the values 0 and 1 are working fine.
Is it not supposed to use cross-validation at Boosting or is this a bug ?
PS: cv::ml::DTrees is the superclass of cv::ml::Boost.