Method for encoding and decoding video information, a motion compensated video encoder and a corresponding decoder

• US 6,738,423 B1
• Filed: 01/21/2000
• Issued: 05/18/2004
• Est. Priority Date: 01/21/2000
• Status: Expired due to Term

• Current Assignee: Nokia Technologies Oy (Nokia Corporation)
• Original Assignee: Nokia Mobile Phones UK Limited (Nokia Corporation)
• Inventors: Karczewicz, Marta, Lainema, Jani
• Primary Examiner(s): Rao, Andy
• Assistant Examiner(s): Parsons, Charles E
• Application Number:
  US09/489,543
• Time in Patent Office: 1,579 Days
• Field of Search: 375/240.03, 375/240.16, 375/240.22
• US Class Current: 375/240.03
• CPC Class Codes: H04N 19/105 Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction
• CPC Class Codes: H04N 19/126 Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
• CPC Class Codes: H04N 19/154 Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion use of rate-distortion criteria H04N19/147
• CPC Class Codes: H04N 19/172 the region being a picture, frame or field
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/196 being specially adapted for the computation of encoding parameters, e.g. by averaging previously computed encoding parameters processing of motion vectors H04N19/513
• CPC Class Codes: H04N 19/198 including smoothing of a sequence of encoding parameters, e.g. by averaging, by choice of the maximum, minimum or median value
• CPC Class Codes: H04N 19/46 Embedding additional information in the video signal during the compression process H04N19/517, H04N19/68, H04N19/70 take precedence
• CPC Class Codes: H04N 19/463 by compressing encoding parameters before transmission
• CPC Class Codes: H04N 19/51 Motion estimation or motion compensation
• CPC Class Codes: H04N 19/517 by encoding
• CPC Class Codes: H04N 19/52 by predictive encoding
• CPC Class Codes: H04N 19/527 Global motion vector estimation
• CPC Class Codes: H04N 19/537 Motion estimation other than block-based
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

A motion compensated video coding method which can be applied especially in transfer of video streams using low transmission bit rate is presented. In the motion compensated coding method, the motion of picture elements between a piece of reference video information and a piece of current video information is estimated and then modeled using certain basis function and coefficients. The coefficients are quantized, and the quantizer is selected according to a certain selection criterion, for example, based on a target image quality or on a target transmission bit rate. Preferably the selection criterion is such that it automatically adjust the accuracy with which the motion of picture elements is represented to be related to the accuracy with which the prediction error information is represented. A decoding method, an encoder and a corresponding decoder are also described.

Claims

1. A method for encoding video information, comprising:

estimating the motion of a picture element between a piece of reference video information and a piece of current video information;

representing the motion of said picture element using a set of basis functions and associated motion coefficients thereby forming a model representative of the motion of said picture element,

forming prediction video information for the picture element from the piece of reference video information, using said model representative of the motion of said picture element,

forming prediction error information representative of a difference between said prediction video information for the picture element and video information of the picture element, the prediction error information having a certain accuracy, wherein the method further comprises;

defining a set of motion coefficient quantizers for quantizing the motion coefficients of said model representative of the motion of said picture element;

selecting a motion coefficient quantizer from the set of motion coefficient quantizers, and

quantizing the motion coefficients using the selected motion coefficient quantizer, thereby representing the motion of said picture element with a certain accuracy,

said selected motion coefficient quantizer being selected using a predetermined selection criterion so that the accuracy with which the motion of said picture element is represented is compatible with the accuracy of said prediction error information.

2. A method for encoding video information according to claim 1, wherein the predetermined selection criterion is a target bit rate for transmission of the encoded video information.

3. A method for encoding video information according to claim 1, wherein said prediction video information for the picture element is formed by:

defining a set of inverse motion coefficient quantizers,

determining a selected motion coefficient quantizer using which the motion coefficients are quantized,

performing inverse quantization of the quantized motion coefficients using an inverse motion coefficient quantizer corresponding to the selected motion coefficient quantizer,

determining the motion of the picture element with respect to the piece of reference video information using the inverse quantized motion coefficients and the basis functions,

determining the prediction video information for the picture element using the piece of reference video information and the determined motion of the picture element,

and said prediction error information is formed by;

determining a piece of prediction error video information based on a difference between the prediction video information for the picture element and the piece of current video information,

transforming the piece of prediction error video information, thereby representing it with prediction error coefficients, and

quantizing the prediction error coefficients using a prediction error quantizer.

4. A method for encoding video information according to claim 3, wherein the motion coefficient quantizer is selected based on the prediction error quantizer.

5. A method for encoding video information according to claim 3, in which the prediction error quantizer has a certain quantization interval, wherein a quantization interval of the motion coefficient quantizer is related to the quantization interval of the prediction error quantizer.

6. A method for encoding video information according to claim 1, wherein the predetermined selection criterion is a target image quality.

7. A method for encoding video information according to claim 1, wherein the predetermined selection criterion is the amount of information needed to represent the quantized coefficients.

8. A method for encoding video information according to claim 1, wherein the motion of the picture element is represented using a set of orthogonal basis functions.

9. A method for encoding video information according to claim 8, wherein the motion of the picture element is represented using a set of affine orthogonal basis functions.

10. A method for encoding video information according to claim 8, wherein the motion of the picture element is represented by predicting the motion of the picture element based on the motion of least one neighboring picture element and by determining a refinement motion for the picture element.

11. A method for encoding video information according to claim 10, wherein the refinement motion is represented using a set of affine orthogonal basis functions.

12. A method for encoding video information according to claim 1, further comprising transmitting the quantized motion coefficients to a receiver.

13. A method for encoding video information according to claim 12, further comprising transmitting information specifying the selected motion coefficient quantizer to the receiver.

14. A method for encoding video information according to claim 1, wherein the set of motion coefficient quantizers comprises a number of uniform quantizers each having a different quantization interval.

15. A method for encoding video information according to claim 1, wherein the set of motion coefficient quantizers comprises a number of modified uniform quantizers, each having a different quantization interval.

18. A method for decoding encoded video information according to claim 5, further comprising determining the basis functions using which the motion of the picture elements is represented.

19. A method for decoding encoded video information according to claim 5, wherein the received prediction error information comprises quantized prediction error coefficients representing a piece of prediction error video information, and the method further comprises:

determining a prediction error quantizer using which the prediction error coefficients are quantized,

performing inverse quantization of the quantized prediction error coefficients using an inverse quantizer corresponding to said prediction error quantizer,

determining a decoded piece of prediction error video information using the inverse quantized prediction error coefficients, and

determining a decoded piece of current video information using the prediction video information and the decoded piece of prediction error information.

20. A method for decoding encoded video information according to claim 19, wherein the selected inverse motion coefficient quantizer is determined based on the prediction error quantizer.

16. A method for decoding encoded video information, comprising:

receiving encoded video information comprising quantized motion coefficients representing the motion of a picture element with respect to a piece of reference video information, and prediction error information, the prediction error information having a certain accuracy, said quantized motion coefficients having been quantized in a video encoder using a motion coefficient quantizer selected using a predetermined selection criterion so that the accuracy with which the motion of said picture element is represented is compatible with the accuracy of said prediction error information,

the method further comprising;

defining a set of inverse motion coefficient quantizers for inverse quantizing the quantized motion coefficients,

determining a selected motion coefficient quantizer using which the motion coefficients are quantized,

selecting an inverse motion coefficient quantizer from the set of inverse motion coefficient quantizers, corresponding to the selected motion coefficient quantizer, and

performing inverse quantization of the quantized motion coefficients using the selected inverse motion coefficient quantizer to form inverse quantized motion coefficients.

17. A method for decoding encoded video information according to claim 16, further comprising:

representing the motion of said picture element using the inverse quantized motion coefficients and a set of basis functions, thereby forming a model representative of the motion of said picture element, and

forming prediction video information for the picture element from the piece of reference video information, using said model representative of the motion of said picture element.

21. A method for decoding encoded video information according to claim 16, wherein the selected motion coefficient quantizer is determined from transmitted information relating to a parameter used in encoding the video information.

22. A method for decoding encoded video information according to claim 16, wherein the encoded video information comprises information indicating the selected motion coefficient quantizer.

23. A method for decoding encoded video information according to claim 16, further comprising a step of receiving signalling information indicating the selected motion coefficient quantizer.

24. An encoder for performing motion compensated encoding of video information, comprising:

an input for receiving a piece of current video information,

a memory for storing a piece of reference video information,

a motion field estimator for estimating the motion of a picture element in the piece of current video information using at least the piece of reference video information,

a motion field coder, comprising;

means for representing the motion of said picture element using a set of basis functions and associated motion coefficients, to form a model representative of the motion of said picture element,

first selection means for selecting a motion coefficient quantizer from a set of motion coefficient quantizers, said first selection means having an input for receiving information indicating a selection criterion and an output for outputting information indicating a selected motion coefficient quantizer, and

quantization means for quantizing motion coefficients using a selected motion coefficient quantizer, said quantization means having a first input for receiving information indicating a selected motion coefficient quantizer, a second input for receiving motion coefficients, and an output for outputting quantized motion coefficients, and

a motion compensated predictor comprising;

second selection means for selecting an inverse motion coefficient quantizer from a set of inverse motion coefficient quantizers, said second selection means having an input for receiving information indicating a selection criterion and an output for outputting information indicating a selected inverse motion coefficient quantizer,

inverse quantization means for inversely quantizing quantized motion coefficients using a selected inverse motion coefficient quantizer, said inverse quantization means having a first input for receiving quantized motion coefficients, a second input for receiving information indicating a selected inverse motion coefficient quantizer and an output for outputting inverse quantized motion coefficients, and

means for determining a piece of prediction video information using at least the piece of reference video information and the inverse quantized motion coefficients.

25. An encoder for performing motion compensated encoding of video information, comprising:

a motion estimator for estimating the motion of a picture element between a piece of reference video information and a piece of current video information,

a motion field coder for representing the motion of said picture element using a set of basis functions and associated motion coefficients thereby forming a model representative of the motion of said picture element,

a motion compensated predictor for forming prediction video information for the picture element from the piece of reference video information, using said model representative of the motion of said picture element,

a prediction error coder for forming prediction error information representative of a difference between said prediction video information for the picture element and video information of the picture element, the prediction error information having a certain accuracy,

wherein the encoder further comprises;

means for defining a set of motion coefficient quantizers for quantizing the motion coefficients of said model representative of the motion of said picture element;

means for selecting a motion coefficient quantizer from the set of motion coefficient quantizers, and

means for quantizing the motion coefficients using the selected motion coefficient quantizer, thereby representing the motion of said picture element with a certain accuracy, and said means for selecting a motion coefficient quantizer is arranged to select a motion coefficient quantizer using a predetermined selection criterion so that the accuracy with which the motion of said picture element is represented is compatible with the accuracy of said prediction error information.

26. An encoder according to claim 25, wherein said motion compensated predictor comprises:

means for defining a set of inverse motion coefficient quantizers,

means for determining a selected motion coefficient quantizer using which the motion coefficients are quantized,

means for performing inverse quantization of the quantized motion coefficients using an inverse motion coefficient quantizer corresponding to the selected motion coefficient quantizer,

means for determining the motion of the picture element with respect to the piece of reference video information using the inverse quantized motion coefficients and the basis functions,

means for determining the prediction video information for the picture element using the piece of reference video information and the determined motion of the picture element,

and said prediction error coder comprises;

means for determining a piece of prediction error video information based on a difference between the prediction video information for the picture element and the piece of current video information,

means for transforming the piece of prediction error video information, thereby representing it with prediction error coefficients, and

means for quantizing the prediction error coefficients using a prediction error quantizer.

27. An encoder according to claim 26, wherein said means for selecting a motion coefficient quantizer is arranged to select a motion coefficient quantizer based on the prediction error quantizer.

28. An encoder according to claim 26, in which the prediction error quantizer has a certain quantization interval, wherein said means for selecting a motion coefficient quantizer is arranged to select a motion coefficient quantizer in dependence on the quantization interval of the prediction error quantizer.

29. An encoder according to claim 25, wherein the predetermined selection criterion is indicative of a target image quality.

30. An encoder according to claim 25, wherein the predetermined selection criterion is indicative of a target bit rate for transmission of the encoded video information.

31. A decoder for the decoding encoded video information, comprising:

a memory for storing a piece of reference video information,

an input for receiving quantized motion coefficients, and

a motion compensated predictor comprising;

selection means for selecting an inverse motion coefficient quantizer from a set of inverse motion coefficient quantizers, said selection means having an input for receiving information indicating a selection criterion and an output for outputting information indicating a selected inverse motion coefficient quantizer,

inverse quantization means for inversely quantizing quantized motion coefficients using a selected inverse motion coefficient quantizer, said inverse quantization means having a first input for receiving quantized motion coefficients, a second input for receiving information indicating a selected inverse motion coefficient quantizer and an output for outputting inverse quantized motion coefficients, and

prediction means for determining a piece of prediction video information using at least the piece of reference video information and the inverse quantized motion coefficients.

32. A decoder for decoding encoded video information, comprising:

an input for receiving encoded video information comprising quantized motion coefficients representing the motion of a picture element with respect to a piece of reference video information, and prediction error information, the prediction error information having a certain accuracy, said quantized motion coefficients having been quantized in a video encoder using a motion coefficient quantizer selected using a predetermined selection criterion so that the accuracy with which the motion of said picture element is represented is compatible with the accuracy of said prediction error information,

the video decoder further comprising;

means for defining a set of inverse motion coefficient quantizers for inverse quantizing the quantized motion coefficients,

means for determining a selected motion coefficient quantizer using which the motion coefficients are quantized,

means for selecting an inverse motion coefficient quantizer from the set of inverse motion coefficient quantizers, corresponding to the selected motion coefficient quantizer, and

means for performing inverse quantization of the quantized motion coefficients using the selected inverse motion coefficient quantizer to form inverse quantized motion coefficients.

33. A decoder according to claim 32, wherein said encoded video information comprises information indicating said predetermined selection criterion.

34. A decoder according to claim 32, further comprising means for receiving signalling information indicating said selection criterion.

35. A decoder for performing the decoding of encoded video information according to claim 32, wherein said prediction error information comprises quantized prediction error coefficients, the decoder further comprising:

means for determining a prediction error quantizer used in the quantization of the prediction error coefficients and a corresponding inverse prediction error quantizer, and

second inverse quantization means for inversely quantizing the quantized prediction error coefficients using the inverse prediction error quantizer, said second inverse quantization means having an output for outputting information indicating the prediction error quantizer,

wherein said means for selecting an inverse motion coefficient quantizer is arranged to select an inverse motion coefficient quantizer responsive to the information indicating the prediction error quantizer.

36. A computer program element for performing motion compensated encoding of video information, comprising computer code for causing a programmable processor means to:

receive a piece of current video information,

store a piece of reference video information,

estimate the motion of a picture element in the piece of current video information using at least the piece of reference video information,

represent the motion of said picture element using a set of basis functions and associated motion coefficients, thereby forming a model representative of the motion of said picture element,

form prediction video information for the picture element from the piece of reference video information, using said model representative of the motion of said picture element,

form prediction error information representative of a difference between said prediction video information for the picture element and video information of the picture element, the prediction error information having a certain accuracy,

the computer program element further comprising computer code for causing a programmable processor means to;

define a set of motion coefficient quantizers for quantizing the motion coefficients of said model representative of the motion of said picture element,

select a motion coefficient quantizer from the set of motion coefficient quantizers, and

quantize the motion coefficients using the selected quantizer, thereby representing the motion of said picture element with a certain accuracy,

said computer code for causing a programmable processor means to select a motion coefficient quantizer being arranged to cause the programmable processor means to select a motion coefficient quantizer using a predetermined selection criterion so that the accuracy with which the motion of said picture element is represented is compatible with the accuracy of said prediction error information.

37. A computer program element according to claim 36, embodied on a computer readable medium.

38. A computer program element for performing the decoding of encoded video information, comprising computer code for causing a programmable processor to:

store a piece of reference video information,

receive encoded video information comprising quantized motion coefficients representing the motion of a picture element with respect to the piece of reference video information, and prediction error information, the prediction error information having a certain accuracy, said quantized motion coefficients having been quantized in a video encoder using a motion coefficient quantizer selected using a predetermined selection criterion so that the accuracy with which the motion of said picture element is represented is compatible with the accuracy of said prediction error information,

the computer program element further comprising computer code for causing a programmable processor means to;

define a set of inverse motion coefficient quantizers for inverse quantizing the quantized motion coefficients,

determine a selected motion coefficient quantizer using which the motion coefficients are quantized,

select an inverse motion coefficient quantizer from the set of inverse motion coefficient quantizers, corresponding to the selected motion coefficient quantizer

perform inverse quantization of the quantized motion coefficients using the selected inverse motion coefficient quantizer.

39. A computer program element according to claim 38, embodied on a computer readable medium.