In GestureRecognitionPipeline::predict_ the preprocessing part is commented out
and thus disabled no matter what preprocessing is set to the pipeline.
Is there a gesture recognition reason for that ? It seems to me that one should apply the same preprocessing for training and predicting but I'm no expert in machine learning. Maybe for getting
a worst case scenario prediction ?
In GestureRecognitionPipeline::predict_ the preprocessing part is commented out and thus disabled no matter what preprocessing is set to the pipeline.
Is there a gesture recognition reason for that ? It seems to me that one should apply the same preprocessing for training and predicting but I'm no expert in machine learning. Maybe for getting a worst case scenario prediction ?