RESPONSIBLE RESEARCHER: Rafael Molina Soriano

TITLE OF THE PROJECT:
High resolution images and video from low resolution compressed video sequences.

KEY WORDS:
High resolution images and video, low resolution video, resolution enhancement, compressed video, motion estimation, Bayesian methods, MPEG-2, MPEG-4.

SUMMARY:
The object of this project is the development of Bayesian methods and techniques to obtain high-resolution images and video from low resolution compressed video sequences.

Such methods and techniques will be integrated in an application that would be used as a post-processing tool for video sequences.

The use of the proposed techniques and methods will reduce both the space and the bandwidth needed to store and transmit video sequences since, once processed, a sequence with similar quality to the obtained using much lower compression levels will be obtained. They will also allow the extraction higher quality frames from low resolution compressed sequences.

The results of the project could be applied immediately in areas as:

1. Frozen image in video where a single frame has not enough quality to be shown or printed,
   
2. High definition television where the algorithms to increase the resolution can solve the conversion between the current TV to the new HDTV problem and,
   
3. Video surveillance where a better image quality is pursued in order to identify interesting objects or people from low-resolution video sequences.