IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 6, NO. 3, MARCH 1997
240 portion of the luminance (Y) component of the SVD-
ﬁltered frame no. 75 (ﬁrst ﬁeld), with
12. (Magniﬁed by a factor
of two). This ﬁgure is almost indistinguishable from the original.
For comparison, it also shows the normalized (from zero to 255)
error between the original frame and the output of the SVD-based
ﬁlter and the normalized error between the original frame and the
output of the 3
3 median ﬁlter. As shown in Fig. 6, the median
ﬁlter extracts both image information and noise, thus causing image
blurring. Hence, such a ﬁlter is not suitable for our application,
which requires near lossless reproduction. In contrast, the SVD ﬁlter
preserves edge details and overall picture ﬁdelity.
We presented a novel noise estimation and ﬁltering algorithm for
still images and video sequences based on the theories of SVD
and data compression. Our experiments show that the technique can
effectively ﬁlter noisy images with no prior knowledge of either
the image or the noise characteristics. This results in increased
compressibility when the ﬁltered data is subsequently processed by
image and video compression schemes, such as JPEG and MPEG. For
still images, comparisons with other ﬁltering schemes, such as But-
terworth ﬁltering and wavelet-based ﬁltering show that SVD-based
ﬁltering is better in preserving edge details. For video sequences,
experiments have shown a 16% improvement in the compression
ratio achieved by nearly lossless MPEG or, equivalently, a visual
quality improvement of 1 dB at the same rate. This scheme can be
used in conjunction with traditional motion-compensated temporal
ﬁltering techniques to further improve the overall performance of a
high-quality video processing system .
The authors thank V. Bhaskaran and C. Herley for discussions,
comments, and suggestions.
 A. Rosenfeld and A. C. Kak,
Digital Picture Processing,
York: Academic, 1982.
 M. K. Ozkan, A. T. Erdem, M. I. Sezan, and A. M. Tekalp, “Efﬁcient
multiframe Wiener restoration of blurred and noisy image sequences,”
IEEE Trans. Image Processing,
vol. 1, pp. 453–476, Oct. 1992.
 H. C. Andrews and C. L. Patterson, “Singular value decompositions
and digital image processing,”
IEEE Trans. Acoust., Speech, Signal
vol. ASSP-24, pp. 26–53, Feb. 1976.
 H.-C. Lee
“Digital image noise suppression method using SVD
block transform,” U.S. Patent 5 010 504, Apr. 1991.
 B. K. Natarajan, “Filtering random noise from deterministic signals
via data compression,”
IEEE Trans. Signal Processing,
vol. 43, pp.
2595–2605, Nov. 1995.
 D. L. Donoho, “De-noising by soft-thresholding,”
IEEE Trans. Inform.
vol. 41, pp. 613–627, May 1995.
 G. H. Golub and C. F. Van Loan,
MD: John Hopkins Univ. Press, 1983.