Apparatus and method for processing video data

A method of processing video signal data, the video signal data having a video frame sequence, comprising the digital processing steps of:

• detecting an object in a video frame sequence, the video frame sequence being from subject video signal data;
• tracking said object through two or more frames of the video frame sequence;
• identifying corresponding elements of said object between two or more video frames, said step of identifying resulting in determined correspondences;
• modeling such determined correspondences and generating a correspondence model;
• resampling pel data corresponding to the object in said video frames, said resampling utilizing said correspondence model;
• segmenting said pel data corresponding to said object from other pel data in said video frame sequence, resulting in segmented object pel data, wherein said segmenting includes temporal integration;
• and
• decomposing said segmented object pel data, said decomposing using a Principal Component Analysis,
• wherein the step of segmenting includes:
• (i) applying block-based motion estimation to the segmented object pel data in multiple video frames,
• (ii) determining finite differences between two or more video frames, and
• (iii) generating an affine motion model from the determined finite differences;
• and
• said modeling includes (i) analyzing the correspondence model using a robust sampling consensus for the solution of the affine motion model, and (ii) analyzing the corresponding elements using a sampling population based on the determined finite differences.