Adaptive Decoding of Mpeg-4 Sprites for Memory-constrained Embedded Systems

Background sprite decoding is an essential part of object-based video coding. The composition and rendering of a final scene involves the placing of individual video objects in a predefined way superimposed on the decoded background image. The MPEG-4 standard includes the decoding algorithm for background image decoding, but this algorithm is not suitable for implementation on a memory-constrained platform. In this paper we present a modification of the decoding algorithm that decodes MPEG-4 sequences while fulfilling the requirements of a memory-constrained multiprocessor system with only 17% extra overhead of computation. Our algorithm reduces the memory cost of such decoding with a factor of four. Additionally, our algorithm offers the possibility of high level data parallelism and consequently contributes to an increase of throughput rate.