Fetching the paper…

Rolling-Unrolling LSTMs for Action Anticipation from First-Person Video · Around