Methods and apparatus for uni-prediction of self-derivation of motion estimation

1. An apparatus, comprising:

a video encoder for encoding at least a portion of a picture, the video encoder comprising a self derivation motion estimator for performing self derivation motion estimation for the portion,

wherein the self derivation motion estimator determines a first motion vector (“MV 0 ”) between the picture and a forward reference picture and a second motion vector (“MV 1 ”) between the picture and a backward reference picture;

wherein the self derivation motion estimator further determines a first difference between MV 0 and a first motion vector predictor, where the first motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimator further determines a second difference between MV 1 and a second motion vector predictor, where the second motion vector predictor is determined based on motion vectors of neighboring blocks, wherein the self derivation motion estimator further selects among uni-directional prediction or bi-directional prediction based on the first difference and the second difference values.

2. The apparatus of claim 1 , wherein a prediction is derived for the portion from at least one of the uni-directional prediction and the bi-directional prediction based on information for reconstructed neighboring blocks with respect to the portion, the information comprising at least one of mode information, motion information, and distortion.

3. The apparatus of claim 1 , wherein the picture is a uni-predictive picture.

4. The apparatus of claim 1 , wherein only a motion vector corresponding to a motion between the picture and a forward reference picture is permitted to be used to encode the portion, another motion vector corresponding to a motion between the picture and another reference picture is restricted from use to encode the portion.

5. The apparatus of claim 1 , wherein a reference index is used to explicitly signal a motion vector used to encode the portion.

6. The apparatus of claim 1 , wherein a combination of implicit signaling and explicit signaling is used to provide an indication to a corresponding decoder of a prediction generated from the uni-directional prediction or the bi-directional prediction and used to encode the portion and which of the uni-directional prediction or the bi-directional prediction is used to generate the prediction.

7. In a video encoder, a method, comprising:

encoding at least a portion of a picture using self derivation motion estimation,

wherein the self derivation motion estimation determines a first motion vector (“MV 0 ”) between the picture and a forward reference picture and a second motion vector (“MV 1 ”) between the picture and a backward reference picture;

wherein the self derivation motion estimation includes determining a first difference between MV 0 and a first motion vector predictor, where the first motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimation includes determining a second difference between MV 1 and a second motion vector predictor, where the second motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimation includes selecting among uni-directional prediction or bi-directional prediction based on the first difference and the second difference values.

8. The method of claim 7 , wherein a prediction is derived for the portion from at least one of the uni-directional prediction and the bi-directional prediction based on information of reconstructed neighboring blocks with respect to the portion, the information comprising at least one of mode information, motion information, and distortion.

9. The method of claim 7 , wherein the picture is a uni-predictive picture.

10. The method of claim 7 , wherein only a motion vector corresponding to a motion between the picture and a forward reference picture is permitted to be used to encode the portion, another motion vector corresponding to a motion between the picture and another reference picture is restricted from use to encode the portion.

11. The method of claim 7 , wherein a reference index is used to explicitly signal a motion vector used to encode the portion.

12. The method of claim 7 , wherein a combination of implicit signaling and explicit signaling is used to provide an indication to a corresponding decoder of a prediction generated from the uni-directional prediction or the bi-directional prediction and used to encode the portion and which of the uni-prediction or the bi-prediction is used to generate the prediction.

13. An apparatus, comprising:

a video decoder for decoding at least a portion of a picture, the video decoder comprising a self derivation motion estimator for performing self derivation motion estimation for the portion,

wherein the self derivation motion estimator determines a first motion vector (“MV 0 ”) between the picture and a forward reference picture and a second motion vector (“MV 1 ”) between the picture and a backward reference picture;

wherein the self derivation motion estimator further determines a first difference between MV 0 and a first motion vector predictor, where the first motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimator further determines a second difference between MV 1 and a second motion vector predictor, where the second motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimator further selects among uni-directional prediction or bi-directional prediction based on the first difference and the second difference values.

14. The apparatus of claim 13 , wherein a prediction is derived for the portion from at least one of the uni-directional prediction and the bi-directional prediction based on information for reconstructed neighboring blocks with respect to the portion, the information comprising at least one of mode information, motion information, and distortion.

15. The apparatus of claim 13 , wherein the picture is a uni-predictive picture.

16. The apparatus of claim 13 , wherein only a motion vector corresponding to a motion between the picture and a forward reference picture is permitted to be used to decode the portion, another motion vector corresponding to a motion between the picture and another reference picture is restricted from use to decode the portion.

17. The apparatus of claim 13 , wherein a motion vector used to encode the portion is explicitly determined from a reference index, the portion being decoded using the motion vector.

18. The apparatus of claim 13 , wherein a prediction generated from the uni-directional prediction or the bi-directional prediction for use to decode the portion and which of the uni-directional prediction or the bi-directional prediction is used to generate the prediction is determined explicitly and implicitly.

19. In a video decoder, a method, comprising:

decoding at least a portion of a picture using self derivation motion estimation,

wherein the self derivation motion estimation determines a first motion vector (“MV 0 ”) between the picture and a forward reference picture and a second motion vector (“MV 1 ”) between the picture and a backward reference picture;

wherein the self derivation motion estimation includes determining a first difference between MV 0 and a first motion vector predictor, where the first motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimation includes determining a second difference between MV 1 and a second motion vector predictor, where the second motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimation includes selecting among uni-directional prediction or bi-directional prediction based on the first difference and the second difference values.

20. The method of claim 19 , wherein a prediction is derived for the portion from at least one of the uni-directional prediction and the bi-directional prediction based on information of reconstructed neighboring blocks with respect to the portion, the information comprising at least one of mode information, motion information, and distortion.

21. The method of claim 19 , wherein the picture is a uni-predictive picture.

22. The method of claim 19 , wherein only a motion vector corresponding to a motion between the picture and a forward reference picture is permitted to be used to decode the portion, another motion vector corresponding to a motion between the picture and another reference picture is restricted from use to decode the portion.

23. The method of claim 19 , wherein a motion vector used to encode the portion is explicitly determined from a reference index, the portion being decoded using the motion vector.

24. The method of claim 19 , wherein a prediction generated from the uni-directional prediction or the bi-directional prediction for use to decode the portion and which of the uni-directional prediction or the bi-directional prediction is used to generate the prediction is determined explicitly and implicitly.

25. A non-transitory computer readable storage media having video signal data encoded thereupon when executed by a processor, performs a method comprising:

decoding, at least a portion of a picture encoded using self derivation motion estimation,

wherein the self derivation motion estimation determines a first motion vector (“MV 0 ”) between the picture and a forward reference picture and a second motion vector (“MV 1 ”) between the picture and a backward reference picture;

wherein the self derivation motion estimation includes determining a first difference between MV 0 and a first motion vector predictor, where the first motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimation includes determining a second difference between MV 1 and a second motion vector predictor, where the second motion vector predictor is determined based on motion vectors of neighboring blocks,

wherein the self derivation motion estimation selects among uni-directional prediction or bi-directional prediction based on the first difference and the second difference values.