Method and device for optimizing the compression of a video stream
A method and device for compressing at least a portion of a video image such that a compressed video stream includes, or does not include, an encoding of a residual depending on whether it is advantageous to do so in terms of costs in rate and in distortion. The size of the video stream and resources needed for decompression are optimized through encoding only the most relevant data in the video stream. If the information added by the residual is insufficiently relevant or useful, the residual is not encoded. Factors associated with the residual are only encoded if use of the residual limits the distortion in a satisfactory manner and if the encoding of the associated factors does not generate too high a rate. Use of the method makes it possible to optimize resources needed for a decoder by optimizing the encoding of the residual.
1. A video compression method executed using an information processing apparatus having one or more processor in a system having ability to compress a portion of an image on a hardware platform, comprising:
determining, via at least one processor, at least one prediction of said image portion;
determining, via at least one processor, a distortion associated with each prediction;
determining a residual, via at least one processor, the residual being a difference between said image portion and said at least one prediction; and
deciding, via at least one processor, whether or not to encode the residual, based on at least the distortion associated with said prediction, wherein said deciding further comprises using said at least one processor for comparing a decision parameter RD 1 and a threshold value RDt, with RD 1 =k +Dp where:
k is an optional term; and
Dp is a distortion value associated with a prediction.
2. A method of claim 1 , wherein, if RD 1 ≦RDt, only every prediction is encoded, and if RD 1 >RDt, every prediction and the residual are encoded.
3. The method according to claim 2 further comprising determining a rate, Rp, associated with every prediction, and calculating the optional term k, with k=Rp* ε, where εis a positive predetermined factor.
4. A video compression method executed using an information processing apparatus having one or more processor in a system having ability to compress a portion of an image on a hardware platform, comprising:
determining, via at least one processor, at least one prediction of said image portion;
determining, via at least one processor, a distortion associated with each prediction;
determining a residual, via at least one processor, the residual being a difference between said image portion and said at least one prediction; and
deciding, via at least one processor, whether or not to encode the residual, based on at least the distortion associated with said prediction, wherein deciding whether or not to encode the residual further comprises comparison of decision parameters RD 1 and RD 2 , calculated as follows:
RD 1= k+Dp;
RD 2= k+Rr*e+Dr
Where:
Rp is the rate associated with each prediction;
Dp is the distortion associated with each prediction;
Rr is the rate associated with the residual;
Dr is the distortion associated with the residual;
e is a positive predetermined factor; and
k is an optional term.
5. A method of claim 4 , wherein during said deciding:
(a) if RD 1 ≦RD 2 , only every prediction is transmitted; and
(b) if RD 1 >RD 2 , every prediction and each residual is transmitted.