Learning to Recognise Spatio-Temporal Interest Points

Learning to Recognise Spatio-Temporal Interest Points

Olusegun T. Oshin (University of Surrey, UK), Andrew Gilbert (University of Surrey, UK), John Illingworth (University of Surrey, UK), and Richard Bowden (University of Surrey, UK)
Copyright: © 2010 |Pages: 17
DOI: 10.4018/978-1-60566-900-7.ch002
OnDemand PDF Download:
No Current Special Offers


In this chapter, we present a generic classifier for detecting spatio-temporal interest points within video, the premise being that, given an interest point detector, we can learn a classifier that duplicates its functionality and which is both accurate and computationally efficient. This means that interest point detection can be achieved independent of the complexity of the original interest point formulation. We extend the naive Bayesian classifier of Randomised Ferns to the spatio-temporal domain and learn classifiers that duplicate the functionality of common spatio-temporal interest point detectors. Results demonstrate accurate reproduction of results with a classifier that can be applied exhaustively to video at frame-rate, without optimisation, in a scanning window approach.
Chapter Preview

Background And Previous Work

This section reviews current methods used in action recognition. First, a short description of relevant spatial interest point detectors is given, followed by examples of local spatio-temporal interest point detectors and methods that make use of them. Global approaches to the task of action recognition are also explored along with a brief evaluation of both local and global approaches. This section also examines common interest point descriptors and methods by which interest point representations are used to achieve action recognition.

Complete Chapter List

Search this Book: