Video compression using color transfer based on motion estimation

Abstract

Many compression techniques were developed in last decade and are being used in many applications like HDTV, Videoconferencing, Videophone, multimedia work stations and mobile image communications. It is also certain that digital video will have a significant economic impact on the computer, telecommunications, and imaging industries. Compression that is obtained by standard compression schemes for color video can be further increased if we can take the advantage of color information of successive images of a scene. The color of the objects in the present frame will be almost similar to the color of it in previous frame. So color can be applied at decoder, even if the information is not known for all frames of a scene. The main objective of this thesis is to propose schemes for different profiles of MPEG-2 which uses color transfer techniques. The proposed schemes are tested with different sequences and are compared with the MPEG-2 coded sequences.