Multivector motion description for region-based very low bit-rate video coding

In the present paper, a new approach to region motion description and estimation is introduced, which results particularly suitable for segmentation-based coding strategies for very low bit-rate video coding. Region motion is described through a variable number of motion vectors (MV's) applied to specific control points. No information about this control points is required to be transmitted, as their determination is based on information available at the decoder. Results show an important net bit-rate saving for QCIF images using this new approach versus the standard translational model. Transmission at rates below 64 kbit/s with very high image quality are achieved.