Deep kernel video approximation for unsupervised action segmentation

Silvia L. Pintea, Jouke Dijkstra

This work focuses on per-video unsupervised action segmentation, which is of interest to applications where storing large datasets is either not possible, or nor permitted. We prop…