Scale Invariant Action Recognition Using Compound Features Mined from Dense Spatio-temporal Corners.

Andrew Gilbert, John Illingworth, Richard Bowden

Browse the full ECCV paper archive.