Video encoding-method

• US 6,654,420 B1
• Filed: 10/27/2000
• Issued: 11/25/2003
• Est. Priority Date: 10/29/1999
• Status: Expired due to Fees

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Snook, Daniel
• Primary Examiner(s): Kelley, Chris
• Assistant Examiner(s): BUGG, GEORGE A
• Application Number:
  US09/698,771
• Time in Patent Office: 1,124 Days
• Field of Search: 375/240.12, 375/240.13, 375/240.14, 375/240.15, 375/240.16, 375/240.24
• US Class Current: 375/240.16
• 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 19/112 according to a given display mode, e.g. for interlaced or progressive display mode
• CPC Class Codes: H04N 19/137 Motion inside a coding unit, e.g. average field, frame or block difference
• CPC Class Codes: H04N 19/56 Motion estimation with initialisation of the vector search, e.g. estimating a good candidate to initiate a search
• CPC Class Codes: H04N 19/577 Motion compensation with bidirectional frame interpolation, i.e. using B-pictures

Abstract

The MPEG-4 standard describes four prediction modes: a direct prediction mode, a bidirectional prediction mode, a forward prediction mode, and a backward prediction mode. These modes may be used for the encoding of a block on the basis of a past reference frame and a future reference frame. Thus, for the encoding of a given block, the invention provides an advantageous strategy leading to the choice of a well adapted prediction mode among the four possible ones. The proposed strategy works in favor of the direct mode and gives criteria to be satisfied when using this particular prediction mode.

Claims

1. A method of encoding a sequence of pictures, each picture being partitioned into non-overlapping blocks of pixels, the method comprising, for a block (Y) belonging to a picture (P1) to be encoded on the basis of a past reference frame (P0) and a future reference frame (P2), hereafter referred to as a block to be encoded, at least the steps of:

deriving for a block (Z) in the future reference frame with the same location as the block to be encoded, an optimum motion vector (MV) on the basis of the past reference frame and a corresponding optimum prediction block (X) in the past reference frame;

deriving the sum of absolute difference between the block in said future reference frame with the same location as the block to be encoded and the optimum prediction block in the past reference frame, hereafter referred to as SADref;

deriving for the block to be encoded, a forward motion vector (MVf) on the basis of the optimum motion vector and a corresponding forward prediction block (W) in the past reference frame

deriving the sum of absolute difference between the block to be encoded and the forward prediction block, hereafter referred to as SADf;

deriving for the block to be encoded, a backward motion vector (MVb) on the basis of the optimum motion vector and a corresponding backward prediction block (V) in the future reference frame;

deriving the sum of absolute differences between the block to be encoded and the backward prediction block, hereafter referred to as SADb;

encoding the block to be encoded according to a direct prediction mode if one of the following conditions is satisfied;

the spatial coordinates of the optimum motion vector are within a given range (I);

the deviation of SADref towards SADb is smaller than a given threshold;

the deviation of SADref towards SADf is smaller than a given threshold.

2. A method of encoding as claimed in claim 1, wherein, when none of the conditions of the last step is fulfilled, the method further comprises at least the steps of:

deriving the minimum of the sum of absolute difference between the block to be encoded and the blocks of the past reference frame, this minimum being hereafter referred to as SADfwd and the corresponding block of the past reference frame being hereafter referred to as forward reference block;

deriving the minimum of the sum of absolute differences between the block to be encoded and the blocks of the future reference frame, this minimum being hereafter referred to as SADbck and the corresponding block of the future reference frame being hereafter referred to as backward reference block;

deriving the sum of absolute differences between the block to be encoded and a block being the average of the future reference block and the past reference block, this sum of absolute differences being hereafter referred to as SADbidir;

when SADfwd is lower than both SADbck and SADbidir, encoding the block to be encoded according to a forward prediction mode;

when SADbck is lower than both SADfwd and SADbidir, encoding the block to be encoded according to a backward prediction mode;

when SADbidir is lower than both SADbck and SADfwd, encoding the block to be encoded according to a forward prediction mode.

3. A filtering device for carrying out a method of encoding a sequence of pictures as claimed in claim 1.

4. A storing medium comprising a software module that stores a set of instructions executable under the control of a computer or a processor and performs at least some of the steps of the encoding method as claimed in claim 1.