VLSI architecture for hierarchical mesh-based motion estimation

Methods for object-based compression and composition of natural and synthetic video content are currently emerging in standards such as MPEG-4 and VRML. This paper shows a novel VLSI architecture for generating content-based video object representation. The architecture uses a novel technique, borrowed from the 3D modeling, to optimize the mesh coding. The architecture generates the mesh nodes location as well as the associated motion vectors. The performance results show that the prototype contributes practical delay, and it can be used in online application and the power consumption shows that it is good enough in mobile application. Moreover, the number of bits used for the coding shows that the architecture is suitable for very low bit rate applications since it reuses the motion vector values.