This thesis presents an integrated approach in modeling, extracting, detecting and classifying deformable contours directly from noisy images. We begin by conducting a case study on regularization, formulation and initialization of the active contour models (snakes). Using minimax principle, we derive a regularization criterion whereby the values can be automatically and implicitly determined along the contour. Furthermore, we formulate a set of energy functionals which yield snakes that contain Hough transform as a special case. Subsequently, we consider the problem of modeling and extracting arbitrary deformable contours from noisy images. We combine a stable, invariant and unique contour model with Markov random field to yield prior distribution that exerts influence over an arbitrary global model while allowing for deformation. Under the Bayesian framework, contour extraction turns into posterior estimation, which is in turn equivalent to energy minimization in a generalized active...

Images of three-dimensional targets are characterized by the target subcomponents visible from a particular target-sensor orientation (target pose), with the image often changing quickly with variable pose. We define a class as a set of contiguous target-sensor orientations over which the associated target image is relatively stationary with aspect, and therefore each target is in general characterized by multiple classes. Our feature parser employs a distinct set of Wiener filters for each class of images, to identify the presence of target subcomponents. The Wiener filters are designed using a KarhunenLoeve expansion. The statistical relationships between the different target subcomponents are modeled via a hidden Markov tree (HMT). The HMT classifier is discussed and example results are presented for forward-looking-infrared (FLIR) imagery of several vehicles. Classification performance is compared with that of several other competing classifiers.

This thesis presents an integrated approach in modeling, extracting, detecting and classifying deformable contours directly from noisy images. We begin by conducting a case study on regularization, formulation and initialization of the active contour models (snakes). Using minimax principle, we derive a regularization criterion whereby the values can be automatically and implicitly determined along the contour. Furthermore, we formulate a set of energy functionals which yield snakes that contain Hough transform as a special case. Subsequently, we consider the problem of modeling and extracting arbitrary deformable contours from noisy images. We combine a stable, invariant and unique contour model with Markov random field to yield prior distribution that exerts influence over an arbitrary global model while allowing for deformation. Under the Bayesian framework, contour extraction turns into posterior estimation, which is in turn equivalent to energy minimization in a generalized active...

We consider the problem of classifying objects using their two dimensional silhouettes in environments generating large aberrant observations (outliers). These may be generated by failures in edge extraction or long tailed imaging noise. We propose a new approach based on circular sub-set autoregressions with robust parameter estimation and novel lag selection procedures. Lag selection is carried out in the test and training sets. The resulting estimates of the spectral function are used in a 'distance' based classifier which substantially out-performs techniques based on the sample covariance function in outlier contaminated data. This algorithm also out-performs a robust analogue of Dubois and Glanz (1986), and is well suited to classification problems where sensitivity to clutter is important. Typical examples are fault identification or the recognition of new objects entering a domain.