Image analysis – Image compression or coding
Reexamination Certificate
2000-01-03
2003-05-27
Do, Anh Hong (Department: 2624)
Image analysis
Image compression or coding
C382S235000
Reexamination Certificate
active
06571015
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a method for compressing image information at a high speed and a method for transferring real-time moving images utilizing the same that makes it possible to generate dynamic compression and transmission sleds.
BACKGROUND OF THE INVENTION
An example of a group of curves referred to as “space filling curves” is Hilbert curves proposed by D. Hilbert in 1891, and a method for compressing image information using scanning of two-dimensional Peano curve that is similar thereto has been proposed (see JP-B-7-22345).
FIG. 1
is a diagram showing examples of conventional two-dimensional Hilbert curves in which FIG.
1
(
a
) shows a case of 2×2 pixels; FIG.
1
(
b
) shows a case of 4×4 pixels; and FIG.
1
(
c
) shows a case of 8×8 pixels.
Such curves are recently applied to various researches covering classification of spectral images, database information searching, image compression, computer holograms and the like because of their high neighborhood retaining properties. In such applications, Hilbert curves are used for processing data distributed in two-dimensional and three-dimensional spaces. In general, an operation of establishing one-to-one correspondence between lattice points in an N-dimensional space (N=2, 3, . . .) and one-dimensional points is referred to as “scanning”, and scanning along a Hilbert curve is referred to as “Hilbert scanning”.
Articles on the prior art relating to methods for compressing image information include:
(1) T. Agui, T. Nagae and M. Nakajima, “Generalized Peano Scans for Arbitrary-sized Arrays”, ITEJ Technical Report Vol. 14, No. 37, pp. 25-30, July 1990
(2) Y. Bandoh and S. Kamata, “A Method of computing a Pseudo-Hilbert Scan Filling in a Rectangular parallelepiped Region”, IEICE Transactions (D-II), Vol. J80-D-II, No. 10, pp. 2864-2867, October, 1997
(3) S. Kamata, A. Perez, E. Kawaguchi, “A Method of Computing Hilbert curves in Two- and Three-Dimensional Spaces”, IEICE Transaction (D-II), Vol. J74-D-II, No. 9, pp. 1217-1226, September 1991.
(4) X. Liu, G. F. Schrack, “An Algorithm for Encoding and Decoding the 3-D Hilbert Order”, IEEE Trans. Image Processing, IP-6, No. 9, pp. 1333-1337, September 1997
However, the use of Hilbert scanning has a problem in that the range of application is limited to regions in regions in an N-dimensional space that satisfies:
N
⏞
2
m
x
…
x
2
m
(
m
is
a
natural
number
)
On the contrary, normal Hilbert scanning is frequently inapplicable to normal images in which rectangles are to be scanned. Under such circumstances, scanning methods for rectangular parallelepiped regions have been proposed as generalization of two-dimensional Hilbert scanning (refer to the above-cited articles (1) and (2) on the prior art). The method according to the article (1) on the prior art still has problems associated with computing time and implementation on a hardware basis because it is a method based on a recursive process.
On the contrary, the method according to the article (2) on the prior art proposed by the inventors is a scanning method at a high speed which is an extended version of the article (3) on the prior art including no recursive process. According to the articles (2) and (3) on the prior art, rules for the generation of Hilbert curves are formed into a table in advance, and it is referred one after another to eliminate recursion.
Such generalization of Hilbert scanning has not been considered for three-dimensional applications because of complicated calculations associated therewith. The method according the article (1) on the prior art is generalization of two-dimensional Hilbert scanning, and no specific method for three-dimensional applications is described. The method according to the article (4) on the prior art remains in a computing method in a cubic region (2
m
×2
m
×2
m
), and nothing is mentioned about a case of rectangular parallel piped region. Therefore, the application of Hilbert scanning to three-dimensional spaces has been strictly limited.
However, the generalization of Hilbert scanning in a three-dimensional space can be carried out similarly to that in a two-dimensional space based on the technique according to the article (2) on the prior art.
Under the above-described circumstances, it is a first object of the invention to provide a method for compressing image information at a high speed and a system for compressing image information in which scanning is carried out on rectangular parallelepiped regions (without being limited to three-dimensional Hilbert scanning) as extension of three-dimensional Hilbert scanning to improve throughput and to achieve cost reduction. The utilization of Hilbert curves for scanning of rectangular parallelepiped regions is referred to as “pseudo Hilbert scanning”.
In conventional systems for transferring and distributing moving images, moving images compressed at a distributing server in advance have been transmitted with quality and at a speed determined at the server.
FIG. 2
shows an example of a processing flow in such a conventional system in which image signals from a television camera are fetched into a computer at constant intervals with means such as video capture, and the fetched images are subjected to image compression to transmit image data. A personal computer at the receiving end decompresses the compressed images to display an image at each update.
However, conventional methods have had the following problems.
(1) In systems such as a television conference, image quality is inevitably reduced because the priority is given to the number of images per unit time. In distribution systems with high image quality, since images are transmitted after they are compressed at the distributing end in advance, image data going to be transmitted represent the past and therefore lose real-time property.
(2) Conventional software compression techniques are not fast enough to perform real-time image compression.
(3) Image compression on a hardware basis is costly because dedicated equipment is required.
(4) Since no consideration is paid to the performance of clients' end and the conditions of networks, it is not possible to send optimum moving images that meet the requirements of clients.
One reason for such situations is the fact that priority has been given to compression ratios and image quality in conventional compression and decompression techniques over the compression speed because they have been intended for retention and reproduction as typical in the case of DVDs. This has resulted in low compression speeds and has necessitated transmission at the maximum speed. It has therefore been common to sacrifice image quality to speed when moving images are transferred in real-time as in television conference systems.
It is therefore a second object of the invention to achieve the transfer of real-time moving images by compressing and transferring fetched images in real-time, to transfer images optimized for the requirements of clients and the conditions of networks in terms of the image size, the number of images, the intervals (time) between the images and the quality of the images and to generate a compression and transfer sled for each client. This makes it possible to perform optimum transfer of moving images to a plurality of clients under different conditions, to achieve the optimum performance of moving image transfer in adaptation to increased speeds of computers in the future and to achieve transfer of moving images in synchronism with the speed of a network.
DISCLOSURE OF THE INVENTION
In order to achieve the first object, according to the invention, there is provided:
[1] a method for compressing image information at a high speed wherein picked moving images comprising three-dimensional information are converted into one-dimensional serial information one by one using a look-up table associated with rectangular parallelepip
Kamata Seiichiro
Matsuo Katsunori
Do Anh Hong
Jordan and Hamburg LLP
Kyushu Electronics System Inc.
LandOfFree
Method for compressing image information at high speed and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for compressing image information at high speed and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for compressing image information at high speed and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3045676