Capturing relative motion and finding modes for action recognition in the wild

Olusegun Oshin, Andrew Gilbert, Richard BowdenCVIU 2014 · DOI · OpenReview

Abstract

Highlights•To recognise actions solely on motion observed in the video.•A quaternary representation of the interest point node test.•Extensive experimental analysis throughout the approach.•State of the art performance on “in the wild” datasets.•Performance gains independent of the features or classification architecture.
BibTeX
@article{oshin2014capturing,
  author = {Olusegun Oshin and Andrew Gilbert and Richard Bowden},
  title = {{Capturing relative motion and finding modes for action recognition in the wild}},
  journal = {Comput. Vis. Image Underst.},
  year = {2014},
  doi = {10.1016/j.cviu.2014.04.005},
}

Loading related content...

Figure