Motion estimator algorithm and system's architecture

Details Motion estimator algorithm and system's architecture Motion estimator algorithm and system's architecture

1998-11-16

2002-09-24

Kelley, Chris (Department: 2613)

Pulse or digital communications

Bandwidth reduction or expansion

Television or motion video signal

C348S699000, C348S097000, C382S236000

Reexamination Certificate

active

06456659

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates to the field of video processing, and more particularly, to an algorithm and an architecture of a motion estimator for implementing video coders compliant with the MPEG-2 standard.
BACKGROUND OF THE INVENTION
The concept of motion estimation is that a set of pixels of a field of a picture may be placed in a position of the subsequent picture obtained by translating the preceding one. These transpositions of objects may expose to the video camera parts that were not visible before, as well as changes of their shape, e.g., zooming. The family of algorithms that identify and associate these portions of images is generally referred to as motion estimation. Such an association allows calculation of the portion of a different image by removing the redundant temporal information making more effective the subsequent process of compression by transformation, quantization and entropic coding.
A typical example of the motion estimation is found in the MPEG-2 standard. A typical block diagram of a video MPEG-2 coder is depicted in FIG.
1
. Such a system is made up of the following functional blocks:
1) Field Ordinator. This blocks is composed of one or several field memories outputting the fields in the coding order required by the MPEG standard. For example, if the input sequence is I B B P B B P etc., the output order will be I P B B P B B . . . I is the Intra-coded picture, and is a field and/or a semi-field containing temporal redundancy. P is the Predicted-picture, and is a field and/or semi-field from which the temporal redundancy with respect to the preceding I or P picture (previously coded/decoded) has been removed. B is the Bidirectionally predicted-picture, and is a field and/or a semi-field whose temporal redundancy with respect to the preceding I and subsequent P (or preceding P and successive P) pictures has been removed. In both cases, the I and P pictures must be considered as already coded/decoded.
Each frame buffer in the format 4:2:0 occupies the following memory space:
standard
720 × 576 × 8 for the luminance (Y) =
3,317,760 bit
PAL
360 × 288 × 8 for the chrominance (U) =
  829,440 bit
360 × 288 × 8 for the chrominance (V) =
  829,440 bit
total Y + U + V =
4,976,640 bit
standard
720 × 480 × 8 for the luminance (Y) =
2,764,800 bit
NTSC
360 × 240 × 8 for the chrominance (U) =
  691,200 bit
360 × 240 × 8 for the chrominance (V) =
  691,200 bit
total Y + U + V =
4,147,200 bit
2) Motion Estimator. This block removes the temporal redundancy from the P and B pictures.
3) DCT. This block implements the discrete cosine transform according to the MPEG-2 standard. The I picture and the error pictures P and B are divided in 8*8 blocks of portions Y, U, V on which the DCT transform is performed.
4) Quantizer (Q). An 8*8 block resulting from the DCT transform is then divided by a quantizing matrix to reduce the magnitude of the DCT coefficients. The information associated with the highest frequencies less visible to human sight tend to be removed. The result is reordered and sent to the successive block.
5) Variable Length Coding (VLC). The coded words output from the quantizer tend to contain a large number of null coefficients, followed by non-null values. The null values preceding the first non-null value are counted. The count figure comprises the first portion of a coded word, the second portion of which represents the non-null coefficients. These paired values tend to assume values more probable than others. The most probable ones are coded with relatively short words (composed of 2, 3 or 4 bits), while the least probable are coded with longer words. Statistically, the number of output bits is less than in the case such methods are not implemented.
6) Multiplexer and Buffer. Data generated by the variable length coder, the quantizing matrices, the motion vectors, and other syntactic elements are assembled for constructing the final syntax considered by the MPEG-2 standard. The resulting bitstream is stored in a memory buffer, the limit size of which is defined by the MPEG-2 standard and cannot be overfilled. The quantizer block Q supports the buffer limit by making the division of the DCT 8*8 blocks dependent upon the filling limit of a memory buffer of the system. The quantizer block Q also supports the buffer limit by making the division of the DCT 8*8 blocks dependent upon the energy of the 8*8 source block taken upstream of the motion estimation and DCT transformation process.
7) Inverse Variable Length Coding (I-VLC). The variable length coding functions specified above are executed in an inverse order.
8) Inverse Quantization (IQ). The words output by the I-VLC block are reordered in the 8*8 block structure, which is multiplied by the same quantizing matrix that was used for its preceding coding.
9) Inverse DCT (I-DCT). The DCT transform function is inverted and applied to the 8*8 block output by the inverse quantization process. A change is made from the domain of spatial frequencies to the pixel domain.
10) Motion Compensation and Storage. At the output of the I-DCT block the following pictures may alternately be present. A decoded I picture or semipicture that must be stored in a respective memory buffer for removing the temporal redundancy with respect to subsequent P and B pictures. A decoded prediction error picture or semipicture P or B that must be summed to the information removed previously during the motion estimation phase. In the case of a P picture, such a resulting sum stored in a dedicated memory buffer is used during the motion estimation process for the successive P pictures and B pictures. These field memories are generally distinct from the field memories used for re-arranging the blocks.
11) Display Unit. This unit converts the pictures from the format 4:2:0 to the format 4:2:2, and generates the interlaced format for displaying the images.
An architecture implementing the above-described coder is shown in
FIG. 2
a.
A distinctive feature is that the field rearrangement block (
1
), the block (
10
) for storing the already reconstructed P and I pictures, and the block (
6
) for storing the bitstream produced by the MPEG-2 coding are integrated in memory devices external to the integrated circuit comprising the core of the coder. The decoder accesses the external memory device through a single interface, suitably managed by an integrated controller.
Furthermore, the preprocessing block converts the received images from the format 4:2:2 to the format 4:2:0 by filtering and sub-sampling the chrominance. The post-processing block implements a reverse function during the decoding and displaying phase of the images. The coding phase also employs the decoding for generating the reference pictures to make operative the motion estimation. For example, the first I picture is coded, decoded, stored (as described in block description
10
), and is used for calculating the prediction error that will be used to code the subsequent P and B pictures. The play-back phase of the data stream previously generated by the coding process uses only the inverse functional blocks (I-VLC, I-Q, I-DCT, etc.), never the direct functional blocks. From this perspective, the coding and the decoding implemented for the subsequent displaying of the images are nonconcurrent processes within the integrated architecture.
A description of the non-exhaustive search motion estimator is provided in the following paragraphs by considering two fields of an image. The following description also applies to semifields of the image. The two fields are Q
1
at the instant t, and the subsequent field Q
2
at the instant t+T, where T is the field period ({fraction (1/25)} sec. for the PAL standard, {fraction (1/30)} sec. for the NTSC standard). Q
1
and Q
2
comprise luminance and chrominance components. Assume that motion estimation is applied onl

Affiliated with

Also associated with

Rating

Motion estimator algorithm and system's architecture does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Motion estimator algorithm and system's architecture, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Motion estimator algorithm and system's architecture will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2861856