METHOD AND SYSTEM FOR MULTI-MODAL COMPONENT-BASED TRACKING OF AN OBJECT USING ROBUST INFORMATION FUSION

摘要

It discloses a system and method for tracking a target. The video sequence comprising a plurality of image frames is received. Samples of the target outer shape distribution - this basic expression is maintained. The target is divided into one or more components. For each component, the sample - in relation to the basic expression is estimated that its position, and inert. The VBDF (Variable-Density Fusion Bandwidth) is applied to each component to determine the most dominant motion. Motion estimation is used to determine the track of the target.