Video data encoder and decoder
First Claim
Patent Images
1. Encoder for performing motion compensated encoding of video data, comprising:
- means (8) for estimating motion vectors describing scene motion displacements of video frame pixels;
means (7) for predicting a video data frame (P) based on a first video data frame (R) and said motion vectors;
means (1) for calculating a prediction error frame (E) based on said predicted video data frame (P) and on a second video data frame (I);
means (2) for encoding said prediction error frame (E) and obtaining prediction error data;
means (4) for transmitting said motion vectors and said prediction error data to a decoder;
said means (2) for encoding said prediction error frame (E) comprising;
means (24) for partitioning said prediction error frame into a plurality of areas; and
determination means for performing at least one of the following operations;
determination of areas in said prediction error frame (E) to be encoded and determination of at least one coding scheme for each area to be encoded of said prediction error frame (E); and
area encoding means (26) for encoding the areas to be encoded of said prediction error frame (E) in accordance with the at least one determination operation of said determination means;
wherein said determination means is adapted to analyse properties of said predicted video data frame (P) and to perform said at least one determination operation based on said analysis.
3 Assignments
0 Petitions
Accused Products
Abstract
This invention relates to compression of prediction error in motion compensated video coding schemes. Prediction frame and motion vectors are used to extract useful information about the properties of the prediction error signal, e.g. its location, directionality, etc. This information does not need to be transmitted because the decoder of the proposed system includes means to extract this information. Information extracted by the encoder and decoder is used to adapt prediction error coding scheme which gives substantial bitrate reductions.
108 Citations
15 Claims
-
1. Encoder for performing motion compensated encoding of video data, comprising:
-
means (8) for estimating motion vectors describing scene motion displacements of video frame pixels;
means (7) for predicting a video data frame (P) based on a first video data frame (R) and said motion vectors;
means (1) for calculating a prediction error frame (E) based on said predicted video data frame (P) and on a second video data frame (I);
means (2) for encoding said prediction error frame (E) and obtaining prediction error data;
means (4) for transmitting said motion vectors and said prediction error data to a decoder;
said means (2) for encoding said prediction error frame (E) comprising;
means (24) for partitioning said prediction error frame into a plurality of areas; and
determination means for performing at least one of the following operations;
determination of areas in said prediction error frame (E) to be encoded and determination of at least one coding scheme for each area to be encoded of said prediction error frame (E); and
area encoding means (26) for encoding the areas to be encoded of said prediction error frame (E) in accordance with the at least one determination operation of said determination means;
wherein said determination means is adapted to analyse properties of said predicted video data frame (P) and to perform said at least one determination operation based on said analysis. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
said determination means comprises means (22) for analysing for each area of said prediction error frame a discontinuity of the image brightness in a corresponding area of the predicted video data frame (P); - and
said determination means (2) is adapted to perform said at least one operation based on said discontinuity analysis.
-
-
3. Encoder according to claim 2, characterized in that said discontinuity analysing means (22) comprises:
-
a linear filter for calculating finite differences (FH) in a vertical direction and a linear filter for calculating finite differences (FV) in a horizontal direction in each area of the predicted video data frame (P); and
means for comparing the sum of absolute values of the outputs of both filters in each area with a first predetermined threshold;
said determination means being adapted to determine that an area of the prediction error frame (E) is not to be encoded if said sum of absolute values calculated for the corresponding area of said predicted video data frame (P) does not exceed said first threshold.
-
-
4. Encoder according to claim 1, characterized in that:
said area encoding means (26) comprises selection means (28, 29) for individually selecting an area encoder (271, . . . , 27n) for each area determined to be encoded among a plurality of different area encoders.
-
5. Encoder according to claim 4, characterized in that:
-
said determination means comprises classifying means (23) for assigning one class of a group of property classes to each area to be encoded of said prediction error frame (E) depending on brightness discontinuity properties of the corresponding area of said predicted video data frame (P); and
said selection means (28, 29) is adapted to select an area encoder (271, . . . , 27n) for each area to be encoded in accordance with the class assigned to the respective area.
-
-
6. Encoder according to claim 5, characterized in that:
said classifying means (23) is adapted to perform a classification based on output values provided by linear directional filters (22) for calculating finite differences (FH) in a vertical direction and finite difference (FV) in a horizontal direction in each area of the predicted video data frame (P) that corresponds to an area to be encoded of said prediction error frame (E).
-
7. Encoder according to claim 5, characterized in that
said classifying means (23) is adapted to perform a classification based on respective brightness discontinuity concentrations in different pixel clusters of each area of the predicted video data frame (P) that corresponds to an area to be encoded of said prediction error frame (E). -
8. Encoder according to claim 6, characterized in that:
said group of classes includes at least one of the following classes;
no discontinuities;
horizontal discontinuities;
vertical discontinuities;
diagonal left slanted discontinuities;
diagonal right slanted discontinuities;
non-directional discontinuities.
-
9. Encoder according to claim 1, characterized in that:
said area encoding means (26) is adapted to encode the data of each area to be encoded of said prediction error frame (E) by means of using selected orthogonal transformations or vector quantization or matching pursuit schemes.
-
10. Encoder according to claim 4, characterized by
each area encoder (271, . . . , 27n) being adapted to comprise a plurality of coding methods (531 to 53k); -
coding method selection means (51, 54, 55) provided for each area encoder (271, . . . , 27n) for selecting a coding method from the respective plurality of coding methods (531 to 53k) in accordance with a criterion related to the performance of the coding methods; and
means (52, 56) for communicating the selection by said coding method selection means (51,54,55) to a video decoder.
-
-
11. Encoder according to claim 10, characterized in that said pluralities of coding methods (531, . . . , 53n) are different for different area encoders (271, . . . , 27n).
-
12. Encoder according to claim 1, characterized by:
-
a prediction error decoder (3) for decoding said prediction error frame (E); and
means (5) for calculating said first video data frame (R) based on said predicted video data frame (P) and said decoded prediction error frame (E*).
-
-
13. Encoder according to claim 1, characterized in that:
-
said determination means (25) comprises means for calculating for each area of said predicted video data frame (P) a magnitude and/or direction of corresponding motion vectors; and
said determination means (22, 25) is adapted to perform said determination based on the calculated magnitude and/or direction of said motion vectors.
-
-
14. Encoder according to claim 13, characterized in that:
said determination means is adapted to determine that an area of the prediction error frame (E) is not to be encoded if said calculated magnitude of said motion vectors corresponding to the area does not exceed a second predetermined threshold.
-
15. Method of motion compensated encoding of video data, comprising the steps:
-
estimating motion vectors describing scene motion displacements of video frame pixels;
predicting a video data frame (P) based on a first video data frame (R) and said motion vectors;
calculating a prediction error frame (E) based on said predicted video data frame (P) and a second video data frame input;
encoding said prediction error frame (E) and obtaining prediction error data;
transmitting said motion vectors and said prediction error data to a decoder;
said step of encoding said prediction error frame (E) comprising;
dividing said prediction error (E) frame into a plurality of areas;
performing at least one of the following determination operations;
determining which of the areas in said prediction error frame (E) are to be encoded and determining a coding scheme for each area to be encoded of said prediction error frame (E); and
encoding each area to be encoded of said prediction error frame (E) in accordance with said determination;
wherein said at least one determination operation is based on analysing properties of said predicted video data frame (P).
-
Specification