The most reliable initial matches are used to reduce the parameter variance on further iterations, minimizing the amount of search required for matching more ambiguous features.

This work proposes a complete sensor-independent visual system that provides robust target motion detection.

Researchers, professional engineers, and graduate students in computer vision, pattern recognition and machine learning, will all find this text an accessible survey of machine learning techniques for vision-based motion analysis.

In summary, in this thesis, we exploit the powerful subspace constraints and develop robust motion segmentation methods in different challenging scenarios where the trajectories are either given as input, or unknown beforehand.

This thesis studies methods to automatically segment moving objects in the presence of camera motion without user interaction. It proposes a fast and robust global motion estimation method oriented to video object segmentation.