A novel subspace tracking using correlation-based projection approximation.

Published in: Signal Process. (2000)

Keyphrases

linear projection
projection matrix
sequential monte carlo methods
null space
real time
dimensionality reduction
principal component analysis

appearance model
convex sets
feature space
particle filter
transformation matrix
object tracking
kalman filter

visual tracking
approximation error
cross correlation
particle filtering
mean shift
high dimensional data
lower dimensional

motion tracking
error bounds
eigendecomposition
closed form
low dimensional
feature selection