I agree that large decision trees are also hard to interpret. The idea in the above paper is that we regularize with respect to average tree path length, based

I agree that large decision trees are also hard to interpret. The idea in the above paper is that we regularize with respect to average tree path length, based on the intuition that smaller trees are more interpretable. (But the issue about pixel thresholds probably still stands.)