Egocentric Video Summarization Based on People Interaction Using Deep Learning

Figure 7

Training sample images, training progress, and sample validation images with predicted labels.