Electrical computers and digital data processing systems: input/ – Input/output data processing – Input/output command process
Reexamination Certificate
2000-05-30
2003-09-23
Gaffin, Jeffrey (Department: 2182)
Electrical computers and digital data processing systems: input/
Input/output data processing
Input/output command process
C710S032000, C348S390100, C348S014160, C348S014160, C348S026000, C348S042000, C348S047000, C348S048000, C348S052000, C348S053000, C348S064000, C348S077000, C348S208400, C345S087000, C345S088000, C345S605000, C375S240000, C375S240120, C375S240080
Reexamination Certificate
active
06625667
ABSTRACT:
The present invention relates to a system for modifying an image displayed on a display device.
A digitized image is a two-dimensional array of picture elements or pixels. The quality of the image is a function of its resolution, which is measured as the number of horizontal and vertical pixels per unit length. For example, in a 640 by 480 display, a video frame consists of over 300,000 pixels, each of which may be defined by one of 16.7 million colors (24-bit). Such an exemplary display typically includes approximately a million bytes of data to represent an image in uncompressed form.
Because of the potentially large amounts of data in each image, it is generally preferable to use an encoding methodology suitable for encoding the uncompressed image data into a compressed form containing less bytes. Encoded data images are generally preferable for use in transmission across a computer network to a display or storage device. The computer network may be, for example, the interconnection between a storage device and an attached display device in a desktop computer, or a data line interconnecting distant computers together. In either case, it is desirable to minimize the number of bytes of data being transmitted across a computer network because low bandwidth networks may not be capable of transmitting a sufficient number of bytes of uncompressed image data fast enough to display images at full frame video rates (60 frame per second). Further, for systems capable of transmitting uncompressed image data fast enough to display images at full frame video rates, it is desirable to free up unneeded bandwidth for signals transmitted through high bandwidth networks.
Images exhibit a high level of pixel-to-pixel correlation which permits mathematical techniques, such as a spatial Fourier transform of the image data, to reduce the number of bytes required to represent the image. By using the spatial Fourier transform, the number of bytes of data is primarily reduced by eliminating high frequency information to which the human eye is not very sensitive. In addition, since the human eye is significantly more sensitive to black and white detail than to color detail, some color information in a picture may be eliminated without significantly degrading picture quality.
There are numerous encoding methods, otherwise referred to as standards, currently being used to encode video images that reduce the number of bytes required to be transmitted across computer networks while simultaneously maintaining image quality.
The H.261 standard is suitable for encoding image data of moving images, such as video, for transmission across computer networks. The H.261 standard is formally known as “Digital Processing of Video Signals—Video Coder/Decoder for Audiovisual Services at 56 to 1536 kbit/s,” American National Standards (ANSI) T1.314.1991, and incorporated herein by reference. A similar standard known as the ITU-T Recommendation H.263, also incorporated herein by reference, discloses a similar standard for video coding for low bitrate communication.
Referring to
FIG. 1
, a H.261 source coder
10
receives digital video
11
in the form of a plurality of nonoverlapping 16×16 pixel blocks at a comparator block
24
. Each 16×16 pixel block of digital video is then further divided into four nonoverlapping 8×8 pixel blocks for calculations.
The source coder
10
has two operational modes. The first operational mode, the intraframe mode, primarily involves a discrete cosine transform (DCT) block
12
transforming each 8×8 pixel block of data to a set of spatial frequency coefficients. The output of the transform block
12
normally is an 8×8 block of data primarily consisting of small numbers and a few large numbers. The spatial frequency coefficients from transform block
12
are inputs to a quantizer block
14
that quantizes the spatial frequency coefficients using a single quantization factor (number). In effect, the quantizer block
14
rounds each spatial frequency coefficient to the nearest multiple of the quantization factor and divides the rounded spatial frequency coefficients by the quantization factor to obtain a data set where the original spatial frequency coefficients are replaced by multiples of the quantization factor. The multiple of the quantization factor for each spatial frequency coefficient is transmitted across a computer network to the decoder (not shown).
The output from the DCT block
12
and quantizer block
14
for large spatial frequency coefficients, which tend to be primarily the lower frequency signal components, is the transmission of a small number representative of the number of multiples of the quantization number. The small spatial frequency coefficients, which tend to be primarily the higher frequency signal components, are normally rounded to zero and thus the quantization multiple is zero. The source coder
10
does not transmit zeros to the decoder. In this manner the number of bytes that need to be transmitted across the computer network to represent an image is significantly reduced.
The second operational mode, the interframe mode, uses a memory characteristic for motion compensation of a slightly moved picture. Each 8×8 set of values from the quantizer block
14
is dequantized by both an inverse quantizer block
16
and an inverse DCT block
18
to obtain an 8×8 block of data that is similar to the original input to the source coder
10
. The picture memory block
20
maintains the 8×8 pixel block of unencoded data until the next 8×8 pixel block representative of the same location in the image is processed by the source coder
10
. A filter block
22
removes some undesirable artifacts, if desired. The comparator block
24
compares the current 8×8 pixel block against the previous 8×8 pixel block stored in the memory block
20
for the same location of the image.
There are three possible outputs from the comparator
24
. First, if the current and previous pixel blocks are the same, then no image data needs to be transmitted to the decoder. Second, if the current and previous pixel blocks are similar, then only the differences need to be transmitted to the decoder. Third, if the current and previous pixel blocks are considerably different then intraframe mode is used to compute the current 8×8 pixel block. For color images, the source coder
10
uses luminance and two color difference components (Y, C
B
and C
R
) The H.261 standard requires that the quantization factor be a single constant number.
The control coding block
26
directs the operation of the source coder
10
. The outputs of the source coder
10
are as follows:
Line
30
a
Flag for INTRA/INTER
Line
30
b
Flag for transmitted or not
Line
30
c
Quantizer indication
Line
30
d
Quantizing index for transform coefficients
Line
30
e
Motion vector
Line
30
f
Switching on/off of the loop filter
“Classified Perceptual Coding With Adaptive Quantization”, IEEE Transactions On Circuits and Systems for Video Technology, Vol. 6, No. 4, August 1996 is similar to the H.261 standard. This method uses a quantization matrix between the output of the DCT block
32
and the input of the quantizer block
14
to allow selected coefficients, such as high frequency coefficients, to be selectively weighted. For example, selected high frequency coefficients could be adjusted to zero so that they are not transmitted to the decoder.
Motion Picture Experts Group 1 (MPEG-1) is another standard suitable for the transmission of moving images across computer networks. MPEG-1, formally known as “Information Technology—Coding Of Moving Pictures and Associated Audio For Digital Storage Media Up To About 1.5 Mbit/s—”, ISO/IEC 11172-2, is herein incorporated by reference. Motion Picture Experts Group 2 (MPEG-2) is yet another standard suitable for the transmission of moving images across computer networks. The MPEG-2, formally known as “Information Technology—Generic Coding Of Moving Pictures and Associated Audio Information: Video”, ISO/IEC 13818-2, is als
Chernoff Vilhauer McClung & Stenzel LLP
Gaffin Jeffrey
Peyton Tammara
Sharp Laboratories of America Inc.
LandOfFree
System for improving efficiency of video encodes does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System for improving efficiency of video encodes, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System for improving efficiency of video encodes will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3003418