Combining Spatio-Temporal Appearance Descriptors and Optical Flow for Human Action Recognition in Video Data | Read Paper on Bytez