Image analysis – Image compression or coding – Shape – icon – or feature-based compression
Reexamination Certificate
2001-02-09
2003-03-25
Johns, Andrew W. (Department: 2621)
Image analysis
Image compression or coding
Shape, icon, or feature-based compression
C382S241000, C382S242000
Reexamination Certificate
active
06539123
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to an image coding/decoding method, an image coding/decoding apparatus and an image coding/decoding program recording medium, and particularly, to image coding with fewer bit number and without deteriorating the picture quality for recording and transmitting image signals having information on the shape of an object so as to process image signals for each object efficiently, and to image decoding for decoding the result of the image coding.
BACKGROUND OF THE INVENTION
The technology for digitizing images into digital image data is dramatically spreading and developing because digital data is easy to record, transmit, edit, copy and transfer. One of the advantages of digitization is the possibility of facilitating data compression. Compression coding is an important technology for data recording and data transmission. The compression coding technology has the established international standards, especially one of which is the MPEG standard that has spread as a general digital standard which can process both video and audio.
The compression coding of digital images processes image data comprising a series of digitized still pictures. In general, the compression coding has the two ways, one of which is an intra-frame coding which compresses a frame (corresponding to a picture) of still picture according to the spatial correlation (the correlation in a frame) while removing redundancy, and the other of which is a inter-frame coding which compresses frames of still pictures which are temporally close to each other, for example, temporally serial frames of pictures, according to the temporal correlation (the correlation among frames) while removing redundancy.
A prior art image coding based on MPEG and the like usually uses the intra-frame coding. If the inter-frame coding is carried out as well, the coded data has a high compression rate. To carry out the inter-frame coding, a decoding process which is a converse process of coding and motion detection and motion compensation processes are carried out to generate a predicted image and then a difference between an image to be coded and the predicted image is calculated using the predicted image as a reference image. Thus the decoding process, and the motion detection and motion compensation processes adversely increase the process load for an apparatus. However, the difference is small when the predicted image has preferable precision, it is possible to increase the coding efficiency by coding the difference more than by coding the image to be coded itself.
As the prediction method employed when the inter-frame coding are carried out, there are some methods, namely a forward prediction based on data which is located at a forward position on time series from the data of an image to be coded in a series of still pictures, a backward prediction based on data which is located at a backward position, and a bidirectional prediction based on data which are located at a forward and backward positions. In general, the intra-frame coding is represented as ‘I’, the forward predictive coding is represented as ‘P’, and the bidirectional predictive coding (including the backward predictive coding) is represented as ‘B’.
When only the intra-frame coding is carried out, or when the forward predictive coding as well as the intra-frame coding are carried out, a series of still pictures to be coded can be processed simply according to time series. As opposed to this, when the backward or bidirectional prediction is carried out, the data which is located at a backward position on time series must be first coded. Therefore, in general, when the inter-frame coding is carried out as well, it is determined in advance which each frame constituting the image data to be coded, an I frame to be subjected to the intra-frame coding or a P frame which can be subjected to the forward predictive coding or a B frame which can be subjected to the bidirectional predictive coding. If the data to be processed is an I frame, the data is subjected to the intra-frame coding. If the data to be processed is a P frame or B frame, the data is subjected to the intra-frame coding or the inter-frame coding. When this coding process is carried out, it is possible to predetermine the ratio of the I frame and the P frame and the B frame according to the purpose of the result of the coding and to the like in the coding apparatus.
FIG. 14
is a diagram for explaining the intra- and inter-frame coding processes of the prior art. In the figure, numerals
1400
to
1406
each designate a frame of image data constituting an image data to be coded. Numerals t
0
to t
6
designate the respective times. The order of the times t0 to t6 indicates the course of time series. In the frames
1400
to
1406
, the frame
1400
is a T frame, the frames
1403
and
1406
are P frames, and the frames
1401
,
1402
,
1404
and
1405
are B frames.
Arrows shown in the figure designate the reference relationships of each frame in the coding process. The frame
1400
which is an I frame is subjected to the intra-frame coding without referring to any other frame. The frame
1403
which is a P frame can be coded referring to the frame
1400
which is located at a forward position on time series. The frame
1401
which is a B frame can be coded referring to the frame
1400
which is located at a forward position on time series and/or the frame
1403
which is located at a backward position on time series.
For that reason, as described above, the frame
1403
must be coded earlier than the frames
1401
and
1402
which are located at a forward position in the frame
1403
, and the I frame and the P frame are given priority to be coded earlier than the B frame. Further, no frames are coded referring to the B frame.
When the bidirectional predictive coding is additionally carried out for the coding process, the apparatus can decide whether the B frame is subjected to the inter-frame coding referring to a forward and backward frames, or the B frame is subjected to either a forward frame, a backward frame, or both frames which are selected as reference frames, or the intra-frame coding is an option as well.
As described above, the inter-frame coding, particularly when the bidirectional predictive coding is carried out as well, contributes to an increase in the process load and requires a storage means which has a large memory capacity for retaining temporally adjacent data. However, prediction with high-level precision makes a difference between a predicted image which is obtained by the prediction and an image to be coded small, whereby coding efficiency can be improved. Thus the coding method is determined according to the performance of an apparatus, the picture quality, the properties of coded data to be required and so on.
On the other hand, a method for coding image signals for each object has often been used in recent years. ISO standardizes these method as MPEG4. In November 1996, what is called the video verification model VM5.0 was worked out. The image signal for each object consists of pixel value signals which indicate brightness and color and are called texture and shape signals which represent the shape of the object. The image signal having this form is being utilized most in the computer graphics technology, and in the field where image sources are created such as the department of producing programs.
FIG.
15
(
a
) to FIG.
15
(
c
) are diagrams for explaining the coding for each object in the prior art. FIG.
16
(
a
) and FIG.
16
(
b
) are diagrams for explaining a signal processing for the coding of each object. FIG.
15
(
a
) shows an example of objects to be coded, which is an image consisting of a background image and a foreground image (a goldfish swimming in a fish tank). FIG.
15
(
b
) shows the foreground (the goldfish). FIG.
15
(
c
) shows the background (water plants and water in the fish tank).
To composite the foreground image and the background image, information which is used for deciding which pixel constituting the composite image
Dang Duy M.
Johns Andrew W.
Matsushita Electric - Industrial Co., Ltd.
Parkhurst & Wendel L.L.P.
LandOfFree
Methods, apparatuses and program recording media for image... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods, apparatuses and program recording media for image..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods, apparatuses and program recording media for image... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3056412