Pulse or digital communications – Bandwidth reduction or expansion
Reexamination Certificate
1998-07-02
2002-04-23
Kelley, Chris (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
C382S243000
Reexamination Certificate
active
06377622
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a coding and decoding of shape information for motional images, and more particularly, to an apparatus and method for scalably coding/decoding shapes by utilizing a scan interleaving method in which a scanning order of vertical and horizontal scannings is applied differently from each other in scalably coding/decoding the shapes.
2. Discussion of Related Art
In encoding scalably shape information it is an art that a plurality of layers having each different resolution are coded and transmitted and decoded. Such art has been required in various art fields. In coding a plurality of layers having each different resolution there are types of a spatial scalability and a temporal scalability.
FIG. 1
shows a conceptual diagram providing the spatial scalable coding method. A base layer has a lower spatial resolution and an enhancement layer has a high spatial resolution. A general coding method is used on the base layer. On an enlargement layer, meanwhile, images on the base layer are up-sampled, that is, the images on the base layer become a screen having a high resolution through an interpolation, and images of size equal to images on the enhancement layer are provided thereon. Thus a coding having a high efficiency can be realized by predicting not only the images from the enhancement layer but also the interpolated images from the enlargement layer. White arrow marks shown in
FIG. 1
indicate a direction of a prediction.
A compression efficiency is high in case that an image of an object is separated from its background then is coded. In order to code only an object separately like this, it needs to encode shape information for separating an object from a background on an image, separately from information of luminance and color signals in each of pixels constituting the object. The shape information is a binary image indicating the interior and the exterior of the object. That is, the interior of object may be as “1” and the exterior of object may be as “0”.
In the scalable coding process, not only color information but also shape information is available to be encoded. In coding to an image of an enhancement layer through a use of an image on a base layer, there are cases of using an intra image and a predicted image. In the shape information scalable coding method, in case of using intra shape information of a base layer, a scan interleaving method is employed representatively.
FIG. 2
is a basic conceptual diagram of the scan interleaving for coding from a base layer to an enhancement layer. Rows of the first, third and fifth are reference scan lines and their values are known to a coder and a decoder. Their values may be pixel values of the base layer. The second and fourth rows are cord scan lines to be coded by the coder. In order to code a pixel value of a cord scan line, pixel values of two upper and lower reference scan lines are scanned. The correlation between the reference scan lines and the cord scan lines is as follows.
There is a case that pixel values of two scan lines and a pixel value of a cord scan line are same. In such a case, shape information has a value of “0” on the background and a value of “1” on the object, or has its opposite value. Most pixels to be coded are included into such a case. There is a case that pixel values on two reference scan lines are different from each other and this case is as a transitional sample, in which the pixel values should be coded. A position which the transitional sample is generated can be noted through the reference scan lines, thus the position information does not need to be transmitted. There is also a case that pixel values on two reference scan lines are same as each other meantime are different from a pixel value on a cord scan line. Such a case is as an exceptional sample, in which position information of the pixel that the exceptional sample is generated should be transmitted, meantime a pixel value does not need to be transmitted since it has a value opposite against pixel values of reference lines.
Accordingly, in employing a scan interleaving method in order for coding a base layer to an enhancement layer, two kinds of data, a transitional sample data (hereinafter, referred to as “TSD”) and an exceptional sample data (ESD) should be coded.
Referring to FIGS.
3
(
a
) and
3
(
b
), it shows a resolution of a base layer and an enhancement layer respectively. The enhancement layer has twice higher resolution all in the width and length directions than the base layer. If the base layer has a resolution of M×N, the enhancement layer has a resolution of (2×M)×(2×N). A pixel position of the base layer is constructed so that it may have a position value of the right lower of each of 2×2 blocks on the enhancement layer. Parts with one set of lines in FIG.
3
(
b
) become each position of pixels which constitute the base layer. In order to encode the enhancement layer by using the base layer, white blocks in FIG.
3
(
b
) should be coded. At this time, parts with one set of lines are same as the value on the base layer, thus there is no need to code the parts with one set of lines again. The parts with one set of lines are utilized when the enhancement layer is coded.
There needs two steps to encode the enhancement layer. FIGS.
4
(
a
) to
4
(
d
) represent vertical and horizontal scannings to code the enhancement layer.
As shown in FIG.
4
(
a
), first, left and right pixel values of the base layer as pixels with one set of lines are used as reference values in coding parts with two sets of lines of the enhancement layer. As shown in FIG.
4
(
b
), next, upper and lower pixel values of pixels shown as parts with one set of lines and encoded with the base layer at the first step are utilized as reference values to encode parts with two sets of lines on the enhancement layer.
In another method, as shown in FIG.
4
(
c
), upper and lower pixel values of the base layer shown as parts with one set of lines are used as reference values in order to encode parts with two sets of lines of the enhancement layer. Then, as shown in FIG.
4
(
d
), left and right pixel values of pixels shown as parts with one set of lines and coded with the base layer at the first step are utilized as reference values in order to encode parts with two sets of lines of the enhancement layer.
In two steps for coding an enhancement layer, the method for using left and right pixel values as a reference value is a horizontal scanning, and the method for using upper and lower pixel values as a reference value is a vertical scanning.
The scan interleaving method has two scanning steps as afore-mentioned. In its scanning order there may be also two methods that the vertical scanning is first progressed and the horizontal scanning is next progressed, and the horizontal scanning is first progressed then the vertical scanning is next progressed. At present, one of such two scanning methods in the scan interleaving method is selected and used.
In using the scan interleaving method, two kinds of data, TSD and ESD, should be coded. A main cause for deciding coding information quantity in the scan interleaving method is influenced by generation quantity of the TSD and ESD. That is to say, the smaller the TSD and ESD are generated, the smaller the coding information quantity becomes. By the way, the generation quantity of the TSD and ESD in coding may become different according to a shape of shape information by an order of the vertical and horizontal scannings. To code the enhancement layer, as shown in FIG.
4
(
a
) to
4
(
d
), the scanning should be progressed two times as the vertical and horizontal scannings. In such scannings, the number of coding pixels in the second scanning is more twice than the first scanning regardless an order of the vertical and horizontal scannings. Its reason is that pixels of the enhancement layer are coded in the first scanning referring to pixels of the base layer which are positioned skipping over each one
Chun Sung-Moon
Chung Jae-Won
Kim Jong-Deuk
Moon Joo-Hee
Hyundai Electronics Ind. Co. Ltd.
Kelley Chris
Lackenbach Siegel Marzullo Aronson & Greenspan
Philippe Grims
LandOfFree
Method and apparatus for coding/decoding scalable shapes by... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for coding/decoding scalable shapes by..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for coding/decoding scalable shapes by... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2872692