Date of Award
Master of Science (MS)
Birchfield, Stanley T
The problem of tracking an object in an image sequence involves challenges like translation, rotation, scaling, varying ambient light and occlusions. A model of an object is built off-line by making a training set with images of the object with different poses. A dimensionality reduction technique is used to capture the variations in the training images. This gives a low-dimensional representation of the data. Isometric feature mapping is the dimensionality reduction technique used to capture the true degrees of freedom in the data. Once the data is reduced to low-dimensions it forms a part of the state-vector of the object. Tracking is done in the Bayesian framework. Particle filters track the object in presence of nonlinearity and non-Gaussianity. The focus of this thesis is the problem of tracking a person's head and also estimating its pose using Isometric feature mapping for dimensionality reduction and particle filter for tracking.
Rane, Nikhil, "ISOMAP TRACKING WITH PARTICLE FILTER" (2007). All Theses. 96.