Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-02-09
2001-12-25
Britton, Howard (Department: 2713)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C375S240240, C375S240270
Reexamination Certificate
active
06333948
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a video coding apparatus and a video coding method for compression-coding a video signal in high efficiency and transmitting the same, and more particularly, it relates to refresh of video coding for eliminating influence caused by a transmission error in a reproduced picture.
2. Description of the Background Art
As the field of application of the technique of compression coding a video signal in high efficiency and transmitting the same, a visual telephone or a video conference shown in
FIG. 34A
is general. Further, application to a system shown in
FIG. 34B
for transmitting a video signal through digital radio communication utilizing a transmission path of wireless LAN for monitoring a danger point or transmitting a picture between mobiles, and application to picture distribution utilizing Internet shown in
FIG. 34C
are expected.
In the case of wireless monitoring, it is indispensable to compression-code the picture signal in high efficiency for efficiently utilizing the frequency band in view of effective use of wireless resources, and the communication quality represented by the error rate lowers by at least two or three digits as compared with a cable system. Therefore, the wireless monitoring is readily influenced by an error when performing high-efficiency compression coding omitting redundant information, and hence it is indispensable to improve error resistance by devising a refresh method or the like as described later. Dissimilarly to the visual telephone, transmission of picture information is generally unidirectional from a camera side to a monitor side in the case of monitoring, and the range of utilization rather widens when the refresh functions through a unidirectional transmission path.
A portable visual telephone employing digital radio communication such as PHS (Personal Handyphone System) is also assumed. In this case, a bidirectional transmission path can be ensured although the quality is inferior as compared with the cable system.
Also in the case of Internet, the Internet network presupposes the best effort, there is no guarantee on the quality of a packet waste rate or the like, and it is indispensable to consider effective utilization of resources and processing against packet loss. In the case of Internet, a bidirectional communication path is ensured and hence a method presupposing bidirectionality can be utilized in 1:1 transmission system. However, in the case of a broadcasting type 1:multiple transmission system, there is such a problem that processing of feedback information on the server side concentrates and hence the range of utilization rather widens when the refresh unidirectionally functions. Further, such a case that the server side is storage information is also assumed in the case of Internet, and hence feedback information cannot be utilized since processing of coding has been ended before the communication time.
In general, there has been devised a method of transmitting coded data with an error detection code on the coding side when a feedback path is included in the communication path and, if an error is detected when the decoding side performs error detection of the received data, noticing the coding side this error through the feedback path so that the coding side INTRA-codes and refreshes all coding after the notice of the error. ITU-T recommendation H.261 standardizes this error notice as a screen update request.
With reference to video coding in a video conference or a visual telephone, the conventional video coding method is now described in detail.
Generally in coding of a video conference or visual telephone signal, it is general to employ coding which combines inter-picture coding utilizing frame-to-frame correlation and intra-picture coding with each other along the frame direction. A television image formed by 30 pictures (frames) per second has large correlation along the time axis direction, and if employing pixels on the same position of a screen precedent by one frame for prediction through Inter-frame correlation, it follows that most ideal prediction can be performed when the screen is still. In INTER coding, however, Inter-frame correlation contrarily lowers if there is motion in the screen, resulting in being rather lower even as compared with correlation between adjacent pixels in a field. On the other hand, each pixel of a picture signal per frame has small level change with respect to an adjacent pixel and its correlation is strong. It is assumed that its self correlation function can be approximated by a negative exponential function. At this time, Power spectral density which is Fourier transform of the self correlation function has a property of being maximized at a zero frequency component (i.e., dc component) and monotonously decreasing as the frequency component increases. While Fourier transform is best known as orthogonal transform to a frequency region, the Fourier transform includes complex number calculation and its structure is complicated, and hence it is general to employ two-dimensional DCT (Discrete Cosine Transform) in coding of pictures as substitute orthogonal transform. After a transform coefficient decomposed into frequency components by DCT is quantized to a level zero which is an uncoded transform coefficient (zero value of the coded coefficient) and to a level ±K from a level ±1 which are non-zero values of the coded coefficient taking discrete quantization representative values, run-length coding for coding the number of successive zeros preceding the coded coefficient and Huffman coding for allocating variable length codes in response to the originating rate of the level of the non-zero value of the coded coefficient are performed, whereby video data are compressed.
For example, ITU-T recommendation H.261 applies motion compensation inter-picture coding to a picture having small motion while performing coding shown below on a prediction error between frames. Further, no inter-picture coding is applied to a picture having large motion but the following coding is directly performed on frame pixels.
FIG. 35
shows an encoder and a decoder for video data according to H.261.
As shown in
FIG. 35
, an encoder
116
for video data according to H.261 comprises a subtraction part
107
, a first orthogonal transform part
108
performing two-dimensional cosine transform, a first quantization part
109
, a second inverse quantization part
110
, a second inverse orthogonal transform part
111
, an addition part
112
, a second picture memory
113
for motion compensation, an in-loop filter
114
, a coding control part
115
and selectors
123
and
124
.
On the other hand, a decoder
122
comprises a first inverse quantization part
117
, a first inverse orthogonal transform part
118
, an addition part
119
, a first picture memory
120
for motion compensation, an in-loop filter
121
and a selector
125
.
The encoder
116
calculates by the subtraction part
107
a prediction error between frames by taking the difference between a video input signal previously transformed to CIF (Common Intermediate Format) of 352 by 288 dots and prediction data stored in the second picture memory
113
for motion compensation. At this time, motion in the range of 15 by 15 pixels is motion-compensated by specifying the prediction data as an arbitrary block of 16 by 16 pixels among 16 by 16 pixels around the block. The motion quantity is specified by a two-dimensional motion vector and transmitted to the decoder along with the video data. The decoding side decodes data of the picture memory for motion compensation in a region displaced from a decoding block by this motion vector as prediction data. For such large motion that no motion compensation is effective, INTRA coding with no prediction is selected by the selectors
123
and
124
. The prediction error and the frame pixels are divided into blocks of 8 pixels by 8 lines, and two-dimensional cosine transform is performed on each block in the first or
Kurobe Akio
Masaki Shoichi
Britton Howard
Matsushita Electric - Industrial Co., Ltd.
Wenderoth, Lind & Poncak, L.L.P.
LandOfFree
Video coding apparatus, video coding method and storage... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Video coding apparatus, video coding method and storage..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Video coding apparatus, video coding method and storage... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2590018