Electrical computers: arithmetic processing and calculating – Electrical digital calculating computer – Particular function performed
Reexamination Certificate
1996-05-14
2001-01-30
Mai, Tan V. (Department: 2787)
Electrical computers: arithmetic processing and calculating
Electrical digital calculating computer
Particular function performed
C708S308000
Reexamination Certificate
active
06182102
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to multimedia computing and communication. More particularly, the invention relates to a system and method for implementing the inverse wavelet transform for the digital compression of still images and video signals.
BACKGROUND OF THE INVENTION
Recently, there has been growing demand for multimedia computing and communication. This growing demand has motivated searches for better bandwidth management techniques, including newer and more efficient compression methods.
Today's mainstream compression methods, such as JPEG for still images and MPEG for video, use the Discrete Cosine Transform (DCT). The sinusoidal basis functions of the DCT have infinite support and so each sinusoidal basis function provides perfect frequency resolution but no spatial resolution. At the other extreme, impulse basis functions have infinitesimal support, so each impulse basis function provides perfect spatial resolution but no frequency resolution. However, neither sinusoidal nor impulse basis functions are very well suited for the purposes of image and video compression. Better suited are basis functions which can trade-off between frequency and spatial resolution.
The wavelet basis functions of a wavelet transform are such basis functions in that each wavelet basis function has finite support of a different width. The wider wavelets examine larger regions of the signal and resolve low frequency details accurately, while the narrower wavelets examine a small region of the signal and resolve spatial details accurately. Wavelet-based compression has the potential for better compression ratios and less complexity than sinusoidal-based compression. The potential for wavelet-based compression is illustrated by
FIGS. 19
,
20
, and
21
.
FIG. 19
is an original 8 bits per pixel 512×512 image of “Lena.”
FIG. 20
is a reconstructed image of Lena after JPEG compression at a compression ratio of approximately 40.
FIG. 21
is a reconstructed image of Lena after wavelet compression at a compression ratio of approximately 40 using a preferred embodiment of the present invention.
FIG. 21
appears less “blocky” than
FIG. 20
because of the varying widths of the wavelet basis functions.
For a practical discussion of wavelet-based compression, see, for example, “Compressing Still and Moving Images with Wavelets,” by Michael, L. Hilton, Bjorn D. Jawerth, and Ayan Sengupta, in
Multimedia Systems,
volume 2, number 3 (1994). Another useful article is “Vector Quantization,” by Robert M. Gray, in
IEEE ASSP Magazine,
April 1984. Both the above articles are herein incorporated by reference in their entirety.
Prior systems and methods for inverse wavelet filtering use conventional filters. A conventional filter does not efficiently compute the inverse wavelet transform of an image because it does not take advantage of the fact that its input is an upsampled stream of data.
SUMMARY OF THE INVENTION
The system and method of the present invention provides two implementations of the inverse wavelet transform for use in an image decompression system. Both implementations do not waste computation power on the zero-valued values inserted into the data stream during an upsampling process. The implementation optimized for low-throughput applications toggles between even and odd modes each clock cycle. In even and odd modes, the transformed values are multiplied by the associated even or odd filter coefficients. The implementation optimized for high-throughput applications multiplies the transformed values by the even and odd filter coefficients separately in two sets of multipliers and outputs two different results each clock cycle.
REFERENCES:
patent: 4817025 (1989-03-01), Asai et al.
patent: 5528527 (1996-06-01), Iwata et al.
Hilton, Michael L., et al,Compressing Still and Moving Images with Wavelets, Multimedia Systems, Apr. 18, 1994, vol. 2, No. 3, pp. 1-17.
Gray, Robert M.,Vector Quantization, IEEE ASSP Magazine, Apr. 1984, pp. 4-29, U.S.A.
Lempel Mody
Ramachandran Loganath
Vafai Manoucher
LSI Logic Corporation
Mai Tan V.
LandOfFree
System and method for implementation of inverse wavelet... 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 and method for implementation of inverse wavelet..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for implementation of inverse wavelet... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2482994