I am trying to use Trainable Weka segmentation plugin to separate the particles from the background. I have faced the problem of choosing the features, and I hope someone can help me
When I run Weka I choose a set of settings (available features), select some regions and run the training. After training and plotting the results I can improve the training by adding some regions or removing them. Then I save the data and the classifier to the files and close the program.
Later, I want to train a new classifier with a different set of settings, but with the same dataset (on the same image). The problem is - when I load the dataset, it has the features stored in it, and I can’t adjust the list of features. Is there a way to overcome this?