Advances in Visual Data Compression and Communication

i

i

i

i

i

i

i

i

Part V

Compressive Communication

Compressive sensing (CS) is a theory for sampling. However, when it is described

by a bipartite graph and is also constrained by binary input, it looks very similar to

low-density parity-check (LDPC) code but with two differences. The first one is that

the symbols in CS are generated by arithmetic sum instead of logical exclusive-OR

in LDPC code. Therefore, we called binary input CS arithmetic code. In addition,

when the operation is arithmetic, the bipartite graph is no longer necessary as binary

weights. The second one is that the connections are assigned with weights. In this

part, we will study how to apply compressive sensing to solving communication

problems. We call this research compressive communication. ...