Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1998-03-18
2004-11-09
Kelley, Chris (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240000, C375S240010, C382S238000
Reexamination Certificate
active
06816553
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to the processing of image signals. In particular it relates to the coding of both the interlace video and progressive video signals.
BACKGROUND OF THE INVENTION
A new video source format referred to as progressive video has been proposed in industries such as those for package-type video contents, video devices, etc., in order to improve the picture quality. However, it became apparent that when progressive video is compressed using the present MPEG digital video compression standard, there is a problem with the inconsistency between the quality that is provided and the bit amount that is necessary. In the present invention, a solution is proposed that is very effective with regard to developing the MPEG-
2
video compression standard.
Television signals are essentially the standardized display of optical images in a certain amount of time and vertical direction, with complete digital video signals being obtained by digitalizing the video for each scan line.
FIG.
1
(
a
) shows a regular interlace video format and FIG.
1
(
b
) shows a progressive video format provided with video of high quality and that does not have artifacts due to an interlace, such as a flicker, etc. Progressive video format is a basic approach for displaying optical images in which first of all the optical image is captured, the captured signals (two-dimensional signals) then being scanned with and made into a set of discrete scan lines. The time period for capturing the image is {fraction (1/60)} sec in countries (includes Japan and U.S.) that apply the NTSC television standard. Each image captured is thus referred to as a frame.
Analog television signals require a wide frequency band-width for broadcasting, so a means that is effective for reducing the analog bandwidth in half by interlacing the scan lines as shown in FIG.
1
(
a
) has been considered. In this technology, each image that was captured is scanned with every other scan line and every other image is complementarily scanned. Each of this type of complementary image is referred to as a field. When interlace video and progressive video are compared, 1 frame of progressive video corresponds to 2 fields of interlace video.
When using interlace, it is necessary to forfeit some degree of quality regardless of the efficiency of the bandwidth. This is indicated best in the explanation for the frequency space of a sampling spectrum shown in FIG.
2
.
FIG. 2
shows the frequency space of progressive and interlace video spectra with the horizonal axis being the time frequency.
In (a), “x” indicates the repetitious frequency of the baseband spectrum in the case of progressive video. In a television of the NTSC system, 480 scanned images are shown on the monitor. The block applied with shading indicates the frequency “width” necessary for restricting the baseband spectrum in order to avoid the aliasing phenomenon that occurs with respect to repetitious spectrum generated in the “x” part.
(b) shows the baseband spectrum without aliashing with respect to interlace video. It is necessary to restrict the baseband spectrum to ½ of the progressive video due to excess repetitious frequency.
(c) shows another method for restricting the bandwidth to ½ with respect to interlace video. This baseband spectrum is preferable to (b) when there is a necessity to maintain a higher resolution during fast motion.
In the case of the interlace video, the bandwidth has to be restricted to ½ in comparison with progressive video in order to avoid the eiriashingu phenomenon. In the two methods shown in FIGS.
2
(
b
) and (
c
), there are many cases of the method in (c) being preferable since the resolution is higher during fast motion. In the case of (b), the fine parts of the image are lost in an unfavorable form as the motion becomes greater and give an impression to the viewer of the image transiently becoming blurry due to the motion. In the case of (c), a higher resolution is maintained even in fast motion, but it is necessary to accordingly forfeit the resolution during standstill. When (a) and (c) are compared, it is readily apparent that progressive video can maintain a vertical frequency bandwidth that is twice that of the interlace video.
The fact that progressive video transmits video signals of (vertical) frequency with a wide bandwidth that is double that of interlace video was explained. However, the interlace video format has been used in ordinary televisions since the start of public broadcasting. The household television presently receives and displays said interlace television signals. On the other hand, most PC monitors display images in the progressive video format for conditions of very high resolution and quality. The problem of whether future televisions should use the progressive video format in order to gain high quality has been a major argument since digital television technology appeared.
MPEG-
1
and MPEG-
2
are presently global standards for broadcast and memory-use digital video compression. Below, the discussion will be restricted to basic P-frame coding for simplification. Extension to the B-frame is easy, so an explanation will be omitted.
In MPEG, interlace video is compressed due to the combination of motion estimation/compression and frequency-space coding that uses DCT (discrete cosine transformation). In order to use this method, the field sequence is paired as shown in FIG.
4
. Each field is alternately scanned as shown in
FIG. 1
, so when two fields are paired—which is composed for one to be the to-be-scanned image at the top in and the other to be the scan image at the bottom—it becomes the same number of scan lines as the entire frame as shown in FIG.
4
. This is actually referred to as the frame of the interlace video.
Next, this frame is divided into blocks with the same dimensions of 16 lines (scan line)×16 picture elements. This is defined as the basic block dimensions with respect to DCT and motion estimation/compensation. The image signal of this unit is referred to as a micro block. If referred to in the original 2 fields, each micro block corresponds to a pair of field blocks with dimensions of 8 lines×16 picture elements as shown in FIG.
3
. In
FIG. 3
, a micro block with dimensions of 16 lines×16 picture elements held within a frame and two corresponding field blocks, referred to as a top field block with dimensions of 8 lines×16 picture elements and a bottom field block with the same dimensions are shown.
In MPEG, motion estimation/compensation is basically executed between the fields and between the frames as shown in FIG.
4
. First of all, the block that best matches within the field that was coded then decoded beforehand with respect to each field block is searched for as shown at the top of the figure. Each of the two field blocks thus has one motion vector. Motion estimation is executed even within the frame block (microblock) as indicated at the bottom half of FIG.
4
. In this case, there is only one vector for each micro block.
After finding these two types of motion vectors, one of these two types of motion estimation is decided based on the condition that the difference between the motion estimation block signal and the coding signal is small. If it is apparent that the frame motion estimation like that shown at the bottom half of
FIG. 4
is more effective, the frame motion estimation is selected with respect to the macroblock. Thereafter, the margin of the motion estimation difference signal is transformed into the coefficient of the frequency space using DCT.
When using DCT, a selection must be made between the DCT used in the frame block and the DCT used in the field block. DCT is used in a block with dimensions of 8 lines×8 picture elements. Therefore, there are 4 blocks (with respect to the brightness component) in a macroblock with dimensions of 16 lines ×16 picture elements. There are two methods for defining a block of 8 lines×8 picture elements as shown in FIG.
5
. It is generally
Brady III W. James
Kelley Chris
Petersen Bret J.
Wong Allen
LandOfFree
Coding method for an image signal does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Coding method for an image signal, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Coding method for an image signal will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3276525