Method and apparatus for multi-hypothesis decoder side-information coding

A computer-implemented method for decompression in digital video compression systems with decoder side-information including previously reconstructed video frames, comprising:

• determining a conditional probability function for each of a plurality of subsets of side-information video frames, wherein each subset of side-information includes pre-stored and received statistical information;
• estimating a plurality of a-posteriori probability functions of a source signal, each posteriori probability function based on a respective one of the conditional probability functions and a respective extrinsic information function determined based on a previously determined a-posteriori probability for a same subset of the side-information;
• evaluating a stopping criterion for decompression given the plurality of a-posteriori probability functions;
• generating the extrinsic information based on the a-posteriori probability functions;
• and
• determining a likelihood threshold for determining a most probable value of a quantized source frame signal based on the a-posteriori probability functions and outputting the quantized source frame, wherein the extrinsic information is generated using the following equation
• E
• 1
• ⁡
• (
• x
• )
• =
• 1
• ∫
• i
• ∈
• {
• 1
• ,
• …
• ⁢
• ,
• K
• }
• ,
• i
• ∈
• 1
• ⁢
• ΠQ
• i
• ⁡
• (
• X
• )
• ⁢
• ⁢
• ⅆ
• X
• ⁢
• ∏
• i
• ∈
• {
• 1
• ,
• …
• ⁢
• ,
• K
• }
• ,
• i
• ∈
• 1
• ⁢
• ⁢
• Q
• i
• ⁡
• (
• X
• )
• ,
• wherein E1(X) is an extrinsic information function, Qi(X) is an a-posteriori probability function, and K is a number of a-posteriori probability functions.

A non-transitory computer readable medium embodying instructions executed by a processor to perform a method for decompression in data compression systems with decoder side-information including a plurality of signals each of which is correlated to a source, the method comprising:

• partitioning a plurality of decoder side-information signals into a plurality of subsets, wherein the decoder side-information signals include pre-stored and received statistical information;
• determining a conditional probability function for each of the subsets of the decoder side-information signals;
• estimating a plurality of a-posteriori probability functions of a source signal, each a-posteriori probability function based on a respective one of the conditional probability functions and extrinsic information;
• evaluating a stopping criterion for decompression given the plurality of a-posteriori probability functions;
• generating the extrinsic information based on the a-posteriori probability functions;
• and
• determining a likelihood threshold for determining a most probable value of a quantized source signal based on the a-posteriori probability functions and outputting the quantized source upon determining to stop decompression, wherein the stopping criterion is computed using the following equation
• SC=(max{maxiε{1, . . . , K}∫(Qi(X)−f(X))2dX,maxi,jε{1, . . . , K}∫(Qi(X)−Qj(X))2dX<θ}),
• wherein SC is the stopping criterion, Qi(X) and Qj(X) are a-posteriori probability functions and f(X) is a marginal probability distribution of the source signal, and K is a number of a-posteriori probability functions.