Image analysis – Image compression or coding – Pyramid – hierarchy – or tree structure
Reexamination Certificate
1999-02-18
2002-03-12
Couso, Jose L. (Department: 2624)
Image analysis
Image compression or coding
Pyramid, hierarchy, or tree structure
C382S232000, C382S239000
Reexamination Certificate
active
06356666
ABSTRACT:
TECHNICAL FIELD
This invention relates to a method of and a device for performing image compression (image coding) and image expansion (image decoding), which are based on lossless wavelet transformations on which attention is focused as a platform of the next-generation image coding system capable of implementing a unified form of lossless coding/lossy coding.
BACKGROUND ART
In an image compressing/expanding device constructed based on conventional linear transformations such as a wavelet transformation, DCT, etc., the implementation of the function of progressively reproducing and displaying image data by fewer numbers of execution memories and fewer numbers of arithmetic operations takes advantage of the fact that an inverse wavelet transformation, a DCT inverse transformation or the like is a linear transformation (linear system). Further, since the arithmetic order of operations necessary for the inverse transformation and addition made upon updating information in a transformation coefficient domain is changeable, the order of these is changed, an inverse transformation is effected only on transformation coefficient information necessary for information updating and the so-obtained result is added to the immediately-preceding reconstructed image data in an image domain, whereby the progressive reconstruction display of the image has been realized.
FIG. 1
is a block diagram of an image expanding device which executes a processing order obtained from the changing of the arithmetic order created based on the contents described in, for example, IEEE Data Compression Conference (DCC-96), April 1996 “Fast Reconstruction of Subband Decomposed Signals for Progressive Transmission”. In the drawing, pre-updating transformation coefficient information b
0
is inversely transformed by an inverse transformer
101
and updated information &Dgr;b
0
is inversely transformed by an inverse transformer
102
. Thereafter, the outputs of both inverse transformers are added together by a transformation coefficient information updater
103
to thereby obtain a reconstructed image (b
0
+&Dgr;b
0
)
104
.
In the progressive reconstruction display of the image by the image compressing/expanding device based on the conventional linear transformations as described above, the inverse transformation is effected only on the transformation coefficient information necessary for information updating by using the linearity of the transformation. The so-obtained result is added to the immediately preceding reconstructed image. Thus, the progressive reconstruction display of the image can be implemented without additionally holding information about transformation coefficients used to reconstruct the immediately preceding reconstructed image. As a result, execution memories could be reduced in number.
Further, the speeding up of processing can be implemented by utilizing a property of the coefficients whose values are zero, increase the number and omitting inverse transformation arithmetic operations on these coefficients when data is reconstructed only from the transformation coefficients necessary for updating.
However, an image compressing/expanding device based on conventional linear transformations has the problem in that a high picture-quality reconstructed image having the image quality of strictly the same level as the original image cannot be obtained due to a quantization error or the like.
While the image compressing/expanding device based on the lossless wavelet transformation can obtain the high picture-quality reconstructed image, the inverse lossless wavelet transformation used as the platform of the image expanding device includes a non-linear arithmetic operation. Therefore, the progressive reconstruction display of the image must be inevitably implemented by updating information in a wavelet transformation coefficient domain and thereafter performing an inverse lossless wavelet transformation on it in order to implement the progressive reconstruction display of the image. Thus, the lossless wavelet transformation coefficients used to generate the pre-updating image, i.e., all the information immediately before the execution of the inverse lossless wavelet transformation must be held to update the displayed image. Therefore, a problem arises in that the size of the execution memory is reduced and a high-speed image progressive reconstruction display cannot be implemented.
Further, a problem arises in that since the number of transformation coefficients zero in value relatively increases if an inverse lossless wavelet transformation is effected on wavelet transformation coefficients comprised of only information necessary for updating, it is quite natural that the number of arithmetic operations at the inverse lossless wavelet transformation can be reduced and the processing can be speeded up by using its relative increase, but the number of the transformation coefficients zero in value relatively decreases because the inverse lossless wavelet transformation is made after the updating of information in a transformation coefficient domain. Thus the speeding up of the processing utilizing the property that many coefficients zero in value are included in the wavelet transformation coefficients indicative of the information necessary for updating cannot be expected. Namely, the image compressing/expanding device based on the linear transformation and the image compressing/expanding device based on the lossless wavelet transformation have both merits and demerits respectively.
With the foregoing problems in view, it is therefore an object of the present invention to provide an image compressing/expanding device based on a lossless wavelet transformation/inverse transformation, which is capable of implementing a progressive reconstruction display of an image while limiting or controlling the consumption of execution memories, thereby obtaining a high picture-quality reconstructed image.
DISCLOSURE OF THE INVENTION
An image compressing/expanding method according to the invention as claimed in claim 1 comprises, on the image compression side, performing quantization and entropy coding on a lossless wavelet transformation coefficient subjected to a lossless wavelet transformation, and on the image expansion side, adding pre-quantization data generated when an inverse lossless wavelet transformation is effected on a lossless wavelet transformation coefficient subjected to entropy decoding and inverse quantization to thereby generate an image, to the next updated data subjected to entropy decoding and inverse quantization, and thereafter performing the inverse lossless wavelet transformation on the result of addition.
Owing to such a construction, the image compressing/expanding method based on the lossless wavelet transformation brings about an effect in that a progressive representation can be implemented and a high picture-quality reconstructed image having image quality of strictly the same level as the original image can be obtained.
An image compressing/expanding method according to the invention as claimed in claim 2 comprises, on the image compression side, performing quantization and entropy coding on each lossless wavelet transformation coefficient subjected a lossless wavelet transformation, and on the image expansion side, effecting an inverse lossless wavelet transformation on each lossless wavelet transformation coefficient subjected to entropy decoding and inverse quantization to thereby generate an image, and generating updated auxiliary information necessary upon updating an image by adding the image generated by the inverse lossless wavelet transformation to the already generated image, from information necessary upon updating the quality of the image by one stage and some of information about each transformation coefficient immediately before the execution of the inverse lossless wavelet transformation.
Owing to such a construction, an effect is brought about in that upon implementation of a progressive reconstruction display, it is unnecessary to hold all the transformation
Atsumi Eiji
Yoshida Masayuki
Couso Jose L.
Do Anh Hong
Mitsubishi Denki & Kabushiki Kaisha
LandOfFree
Image compressing/expanding method and image compression... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image compressing/expanding method and image compression..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image compressing/expanding method and image compression... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2873376