Representation of motion vectors in modified advanced video coding with spatial scalability

The paper describes improved representation of motion vectors in enhancement-layer bitstream of a spatially-scalable modified AVC video encoder. Proposed solution exploits rescaled motion vectors from the base layer for better predictive encoding of motion vectors in the enhancement layer. The base-layer vectors are used as reference for prediction in the enhancement layer only when no good reference is available in the enhancement layer. The low-resolution base-layer bitstream is fully compatible with AVC standard, while few modifications have been done in the enhancement layer bitstream semantics.