1 | initial version |
If "haarcascade" means a cascade trained on Haar features, the answer is of course NO. You need retrain the cascasde for the LBP features at least because for computing new features and choosing the decision ones you have to have a set of images. But it's easy to train LBP cascade if you have the training data that was used to train your existing Haar cascade even if you used haartraining app. For the LBPs you should use traincascade app and follow to the doc.