Text this: An Interpretable Deep Learning-Based Feature Reduction in Video-Based Human Activity Recognition