Estimation theory and tracking of deformable objects
Patricio Antonio Vela, Ibrahima J. Ndiour · 2010
This paper provides a summary of several recently detailed recursive, filtering strategies as applied to the task of tracking deformable objects from a video sequence. Estimation theory provides the foundation for filtering objects based on presumed dynamics and (partial) uncertain measurements. All of the methods seek to arrive at an accurate contour signal in spite of measurement and prediction errors. They differ fundamentally in the state definition and in the estimation approaches.