Instructor Activity Recognition through Deep Spatiotemporal Features and Feedforward Extreme Learning Machines

Figure 3

Some sample video frames from single-view IAVID-I and multiview MuHAVI-uncut and IXMAS activity recognition datasets.