IEEE TRANSACTIONS ON FUZZY SYSTEMS, VOL. 11, NO.

4, AUGUST 2003 429

Noise Reduction by Fuzzy Image Filtering

Dimitri Van De Ville, Member, IEEE, Mike Nachtegael, Dietrich Van der Weken, Etienne E. Kerre,
Wilfried Philips, Member, IEEE, and Ignace Lemahieu, Senior Member, IEEE

Abstract—A new fuzzy filter is presented for the noise reduc- Therefore, this paper presents a new technique for filtering
tion of images corrupted with additive noise. The filter consists of narrow-tailed and medium narrow-tailed noise by a fuzzy
two stages. The first stage computes a fuzzy derivative for eight dif- filter. Two important features are presented: first, the filter
ferent directions. The second stage uses these fuzzy derivatives to
perform fuzzy smoothing by weighting the contributions of neigh- estimates a “fuzzy derivative” in order to be less sensitive to
boring pixel values. Both stages are based on fuzzy rules which local variations due to image structures such as edges; second,
make use of membership functions. The filter can be applied it- the membership functions are adapted accordingly to the noise
eratively to effectively reduce heavy noise. In particular, the shape level to perform “fuzzy smoothing.”
of the membership functions is adapted according to the remaining The construction of the fuzzy filter is explained in Section II.
noise level after each iteration, making use of the distribution of the
homogeneity in the image. A statistical model for the noise distribu- For each pixel that is processed, the first stage computes a fuzzy
tion can be incorporated to relate the homogeneity to the adapta- derivative. Second, a set of 16 fuzzy rules is fired to determine
tion scheme of the membership functions. Experimental results are a correction term. These rules make use of the fuzzy derivative
obtained to show the feasibility of the proposed approach. These as input. Fuzzy sets are employed to represent the properties
results are also compared to other filters by numerical measures , , and . While the membership func-
and visual inspection.
tions for and are fixed, the membership
Index Terms—Additive noise, edge preserving filtering, fuzzy function for is adapted after each iteration. The adapta-
image filtering, noise reduction. tion scheme is extensively explained in Section III and can be
combined with a statistical model for the noise. In Section IV,
I. INTRODUCTION we present several experimental results. These results are dis-
cussed in detail, and are compared to those obtained by other

T HE application of fuzzy techniques in image processing is

a promising research field [1]. Fuzzy techniques have al-
ready been applied in several domains of image processing (e.g.,
filters. Some final conclusions are drawn in Section V.

II. FUZZY FILTER

filtering, interpolation [2], and morphology [3], [4]), and have
numerous practical applications (e.g., in industrial and medical The general idea behind the filter is to average a pixel using
image processing [5], [6]). other pixel values from its neighborhood, but simultaneously
In this paper, we will focus on fuzzy techniques for image to take care of important image structures such as edges.1 The
filtering. Already several fuzzy filters for noise reduction main concern of the proposed filter is to distinguish between
have been developed, e.g., the well-known FIRE-filter from local variations due to noise and due to image structure.
[7]–[9], the weighted fuzzy mean filter from [10] and [11], In order to accomplish this, for each pixel we derive a value
and the iterative fuzzy control based filter from [12]. Most that expresses the degree in which the derivative in a certain
fuzzy techniques in image noise reduction mainly deal with direction is small. Such a value is derived for each direction
fat-tailed noise like impulse noise. These fuzzy filters are able corresponding to the neighboring pixels of the processed pixel
to outperform rank-order filter schemes (such as the median by a fuzzy rule (Section II-A).
filter). Nevertheless, most fuzzy techniques are not specif- The further construction of the filter is then based on the ob-
ically designed for Gaussian(-like) noise or do not produce servation that a small fuzzy derivative most likely is caused by
convincing results when applied to handle this type of noise. noise, while a large fuzzy derivative most likely is caused by
an edge in the image. Consequently, for each direction we will
apply two fuzzy rules that take this observation into account
Manuscript received November 24, 2001; revised June 27, 2002 and (and thus distinguish between local variations due to noise and
November 13, 2002. The work of D. Van De Ville was supported by the Fund due to image structure), and that determine the contribution of
for Scientific Research—Flanders (Belgium) through a mandate of Research
Assistant. The work of M. Nachtegael and D. Van der Weken was supported by the neighboring pixel values. The result of these rules (16 in
the GOA-project 12.0513.98 by Ghent University, Belgium. total) is defuzzified and a “correction term” is obtained for the
D. Van De Ville was with Ghent University, Belgium. He is currently with the processed pixel value (Section II-B).
Biomedical Imaging Group, the Swiss Federal Institute of Technology Lausanne
(EPFL), CH1015 Lausanne, Switzerland.
M. Nachtegael, D. Van der Weken, and E. Kerre are with the Fuzziness and A. Fuzzy Derivative Estimation
Uncertainty Research Modeling Unit, the Department of Applied Mathematics Estimating derivatives and filtering can be seen as a
and Computer Science, Ghent University, B9000 Ghent, Belgium.
W. Philips is with the Department of Telecommunications and Information chicken-and-egg problem; for filtering we want a good indica-
Processing, Ghent University, B9000 Ghent, Belgium. tion of the edges, while to find these edges we need filtering.
I. Lemahieu is with the Department of Electronics and Information Systems,
Ghent University, B9000 Ghent, Belgium. 1Other fuzzy filters, such as the smoothing fuzzy control based filter [12], also
Digital Object Identifier 10.1109/TFUZZ.2003.814830 take care of edges, but after instead of simultaneous with the noise filtering.

1063-6706/03$17.00 © 2003 IEEE

430 IEEE TRANSACTIONS ON FUZZY SYSTEMS, VOL. 11, NO. 4, AUGUST 2003

(a)

(a)

(b)

(c)
(b)
Fig. 2. Membership functions (a) small, (b) positive, and (c) negative.
Fig. 1. (a) Neighborhood of a central pixel (x; y ). (b) Pixel values indicated
in gray are used to compute the “fuzzy derivative” of the central pixel (x; y )
for the NW -direction.

TABLE I
PIXELS INVOLVED TO CALCULATE THE FUZZY
DERIVATIVES IN EACH DIRECTION

Fig. 3. Relationship between the homogeneity  and the noise level 

2
empirically measured by patches of size 9 9 (N = 9). The accuracy of  is
shown by the standard deviation of  itself.

derivative value will be large, but also derivative

In our approach, we start by looking for the edges. We try to values of neighboring pixels perpendicular to the edge’s direc-
provide a robust estimate by applying fuzzy rules. tion can expected to be large. For example, in the -direction
Consider the neighborhood of a pixel as displayed we can calculate the values ,
in Fig. 1(a). and [see Fig. 1(b)]. The idea is to cancel
A simple derivative at the central pixel po- out the effect of one derivative value which turns out to be high
sition in the direction ( due to noise. Therefore, if two out of three derivative values are
) is defined as the small, it is safe to assume that no edge is present in the con-
difference between the pixel at and its neighbor in the sidered direction. This observation will be taken into account
direction . This derivative value is denoted by . when we formulate the fuzzy rule to calculate the fuzzy deriva-
For example tive values.
In Table I, we give an overview of the pixels we use to cal-
culate the fuzzy derivative for each direction. Each direction
(column 1) corresponds to a fixed position (column 2); the sets
(1)
in column 3 specify which pixels are considered with respect to
Next, the principle of the fuzzy derivative is based on the the central pixel .
following observation. Consider an edge passing through the To compute the value that expresses the degree to which the
neighborhood of a pixel in the direction. The fuzzy derivative in a certain direction is small, we will make
VAN DE VILLE et al.: NOISE REDUCTION BY FUZZY IMAGE FILTERING 431

2
Fig. 5. Histogram of the homogeneity of 9 9-blocks for the “cameraman”
test image. The 20% percentile  of the most homogeneous blocks shifts to
the left as the image is more corrupted, i.e.,  equals 0.96, 0.90, 0.82, and
(a) 0.66 for these cases.

set . These rules are implemented using the minimum to

represent the AND-operator, and the maximum for the OR-oper-
ator. A defuzzification is not needed since the second stage, i.e.,
the fuzzy smoothing, directly uses the membership degrees to
.
The robustness we try to achieve by the fuzzy derivative is by
combining multiple simple derivatives around the pixel and by
making the parameter adaptive. The proper choice of will
be discussed later.

B. Fuzzy Smoothing
To compute the correction term for the processed pixel
value, we use a pair of fuzzy rules for each direction. The idea
(b) behind the rules is the following: if no edge is assumed to be
Fig. 4. Original test images. (a) “Cameraman.” (b) “Boats.” present in a certain direction, the (crisp) derivative value in that
direction can and will be used to compute the correction term.
use of the fuzzy set . The membership function The first part (edge assumption) can be realized by using the
for the property is the following [see Fig. 2(a)]: fuzzy derivative value, for the second part (filtering) we will
have to distinguish between positive and negative values.
For example, let us consider the direction . Using the
(2) values and , we fire the following two
rules, and compute their truthness and :
where is an adaptive parameter (see Section III).
For example, the value of the fuzzy derivative for
the pixel in the -direction is calculated by applying
the following rule:

For the properties and , we also use linear

membership functions [see Fig. 2(b) and (c)]. Again, we im-
plement the AND-operator and OR-operator by respectively the
minimum and maximum. This can be done for each direction
.
The final step in the computation of the fuzzy filter is the
(3)
defuzzification. We are interested in obtaining a correction term
Eight such rules are applied, each computing the degree of mem- , which can be added to the pixel value of location .
bership of the fuzzy derivatives , , to the Therefore, the truthness of the rules and , (so
432 IEEE TRANSACTIONS ON FUZZY SYSTEMS, VOL. 11, NO. 4, AUGUST 2003

(a) (a)

(b)
(b)
Fig. 7. MSE (mean squared error) for (a) “cameraman” and (b) “boats.” ( =
Fig. 6. MSE (mean squared error) for (a) “cameraman” and (b) “boats.” ( = 20.)
5).

for all directions) are aggregated by computing and rescaling

the mean truthness as follows: computed, and the hypothesis comes in: the percentile of the
most homogeneous blocks is determined. We assume this per-
(4) centile is a measure for the homogeneity of “typical” noise in
the image. Using a statistical model for the noise distribution,
we will show that there is a linear relationship between the ho-
where contains the directions and represents the number mogeneity and the standard deviation.
of gray levels. So, each directions contributes to the correction Assume noise samples , , independently
term . and identically distributed, with a probability density func-
tion (PDF) and cumulative density function (CDF)
III. ADAPTIVE THRESHOLD SELECTION . Since a change of the standard deviation rescales
Instead of making use of larger windows to obtain better re- the PDF, the maximum and minimum of samples are scaled
sults for heavier noise, we choose to apply the filter iteratively. the same way. This establishes a linear relationship between
The shape of the membership function is adapted each the homogeneity and the standard deviation. This can also be
iteration according to an estimate of the (remaining) amount of derived more formally. We assume the expectation value
noise. The method assumes that a percentage of the image can to be zero, and the variance to be . If we scale the
be considered as homogeneous and as such can be used to esti- PDF with a factor , we can obtain the following general result:
mate the noise density.
We start by dividing the image in small nonoverlap- (6)
ping blocks. For each block , we compute a rough measure
(7)
for the homogeneity of this block by considering the maximum
and minimum pixel value Next, we define the maximum and minimum of the samples
as
(5)

This measure is commonly used in the context of fuzzy image

processing [13]. Next, a histogram of the homogeneity values is
VAN DE VILLE et al.: NOISE REDUCTION BY FUZZY IMAGE FILTERING 433

for which we can derive the CDFs as

Using (6) and (7), we can show that and are

scaled according to , i.e.,

Therefore, there is a linear relationship between the (expectation

value of the) homogeneity of the samples and the standard
deviation (a)

(8)

where is the slope. Note the correspondence of to

.
The value of the factor can be determined empirically.
A large number of synthetic patches (of size ) are gen-
erated. Each patch consists of noise with the presumed distri-
bution. The effective noise level and the homogeneity of each
patch are measured. The mean value and standard deviation are
calculated for the whole test set. This experiment is done for
several noise levels, resulting in the relationship between the
homogeneity and the noise level. Fig. 3 shows the result for the
case of and 200 experiments for several noise levels. The
errorbars indicate the standard deviation on the noise level esti- (b)
mates.2 We carried out this experiment for Gaussian, Laplacian, Fig. 8. Parameter K for “boats.” (a)  = 5. (b)  = 20.
and uniform noise, obtaining a of, respectively, 52.1, 41.8,
and 75.2.
Next, we use the hypothesis that at least a percentage of the
blocks were originally homogeneous (before the noise degra-
dation). The histogram of the homogeneity of the blocks in the
image is computed, and a percentile of the most homogeneous
blocks is obtained. The value of this percentile is related to
our estimate for the noise variance by the linear relationship
we derived before. A final amplification factor (see later for
its choice) is used to get the parameter
(9)
(a) (b)
This scheme is applied before each iteration to obtain the pa-
rameter , which determines the shape of the membership func-
tion .
Compared to the direct calculation of the variance of (a part
of) the image, the current scheme distinguishes between blocks
containing mainly noise and blocks containing both image
structure and noise. This is done by the sorting operation of the
histogram operation on the homogeneity values. As a result,
the estimate of the noise variance is based on smooth blocks
only, for as long as the initial hypothesis remains true.
(c) (d)
IV. RESULTS
Fig. 9. (a) “Cameraman” with additive gaussian noise ( = 5). (b) After
The proposed filter is applied to grayscale test images (8-bit, 2
Wiener filtering (3 3). (c) After fuzzy mean (FM). (d) After proposed fuzzy
), after adding Gaussian noise of different levels. Such a filter ( = 1).
procedure allows us to compare and evaluate the filtered image
2We also note that the standard deviation of the estimated homogeneity is against the original one. Fig. 4 shows two representative test
very low. images: “cameraman” and “boats.”
434 IEEE TRANSACTIONS ON FUZZY SYSTEMS, VOL. 11, NO. 4, AUGUST 2003

(a) (b)

(a) (b)

(c) (d)
Fig. 11. (a) “Boats” with additive gaussian noise ( = 20). (b) After Wiener
2
filtering (3 3). (c) After AWFM2. (d) After proposed fuzzy filter ( = 2).
(c) (d)
Fig. 10. Detail images of the results of Fig. 9.

Fig. 5 shows the normalized histogram of the homogeneity

of “cameraman,” for the original image, but also for the image
corrupted with different noise levels, i.e., , , and
. Using the 20% percentile and (8), the estimates for
the noise levels are, respectively, 5.2, 9.4, and 17.7. For these
noise levels our filter is applied using different values for the
amplification factor , namely . To evaluate the
results, we computed the mean squared error (MSE) between (a) (b)
the original image and the filtered image.
Figs. 6 and 7 show a plot of the MSE as function of the
number of iterations for added noise with and .
Notice that for low noise levels (Fig. 6), one iteration is suf-
ficient to efficiently remove the noise. Also, a low amplication
factor gives the best results. The MSE of “cameraman” sur-
prisingly increases with the number of iterations, this is mainly
due the image content, i.e., the grass is very similar to noise and
gets increasingly filtered. For other images, such as “boats,” this
increase does not occur. Therefore, images with low noise levels
and containing fine textures should be treated carefully. (c) (d)
For high noise levels (Fig. 7), the results of “cameraman”
Fig. 12. Detail images of Fig. 11.
are much more stable. A few iterations (3–4) are sufficient to
effectively smooth out the noise. Also, a somewhat higher value
of gives better results. curves, could also be determined using the estimate : a high
Fig. 8 shows the parameter for the “boats” test image. noise level corresponds to a higher value of , while a low noise
Since depends on the estimate for the remaining noise level level corresponds to a lower value of .
, we expect this curve to decrease as iterations go on. Based We also compared our fuzzy filter with several other filter
on an estimate for the “natural” or “acceptable” amount of noise techniques: the mean filter, the adaptive Wiener filter [14], fuzzy
(depending on the application), we could use the estimate of median (FM) [15], the adaptive weighted fuzzy mean (AWFM1
as a stop criterion as it gets sufficiently low. Another possible and AWFM2) [10], [11], the iterative fuzzy filter (IFC), modi-
stop criterion could be when the change with respect to the fied iterative fuzzy filter (MIFC), and extended iterative fuzzy
previous iteration is small. filter (EIFC) [12]. Table II summarizes the results we obtained.
The parameter affects the amount of smoothing which is Quite different results are obtained between “cameraman” and
applied by the filter. Based on our observations of the MSE- “boats.” For “cameraman,” the proposed filter performs very
VAN DE VILLE et al.: NOISE REDUCTION BY FUZZY IMAGE FILTERING 435

TABLE II filter is able to preserve the very small details (such as the narrow
RESULTS OF THE NEW FUZZY FILTER FOR THE ropes). On the other hand, the proposed filter gives a more “nat-
TEST IMAGES “CAMERAMAN” AND “BOATS”
ural” image without the “patchy look” of the adaptive Wiener
filter.
Finally, we like to show a practical application of the fuzzy
filter. In particular, this image restoration scheme could be
used to enhance satellite images. Of course, since the original
image is already corrupted by noise, it is not possible to obtain
a numerical measure which indicates how “good” the image is.
Fig. 13 shows the original image and the results after fuzzy fil-
tering with different parameters. Depending on the application
(e.g., visual inspection, segmentation), one could prefer lighter
or heavier filtering (by choosing correspondingly).

V. CONCLUSION
This paper proposed a new fuzzy filter for additive noise re-
duction. Its main feature is that it distinguishes between local
variations due to noise and due to image structures, using a fuzzy
derivative estimation. Fuzzy rules are fired to consider every di-
rection around the processed pixel. Additionally, the shape of
the membership functions is adapted according to the remaining
amount of noise after each iteration. Experimental results show
the feasibility of the new filter and a simple stop criterion. Al-
though its relative simplicity and the straightforward implemen-
tation of the fuzzy operators, the fuzzy filter is able to compete
with state-of-the-art filter techniques for noise reduction. A nu-
merical measure, such as MSE, and visual observation show
convincing results. Finally, the fuzzy filter scheme is sufficiently
simple to enable fast hardware implementations.
(a) (b)

REFERENCES
[1] E. Kerre and M. Nachtegael, Eds., Fuzzy Techniques in Image
Processing. New York: Springer-Verlag, 2000, vol. 52, Studies in
Fuzziness and Soft Computing.
[2] D. Van De Ville, W. Philips, and I. Lemahieu, Fuzzy Techniques in
Image Processing. New York: Springer-Verlag, 2000, vol. 52, Studies
in Fuzziness and Soft Computing, ch. Fuzzy-based motion detection
and its application to de-interlacing, pp. 337–369.
[3] M. Nachtegael and E. E. Kerre, “Connections between binary, gray-scale
and fuzzy mathematical morphologies,” Fuzzy Sets Syst., to be pub-
lished.
[4] , “Decomposing and constructing fuzzy morphological operations
(c) (d) over -cuts: Continuous and discrete case,” IEEE Trans. Fuzzy Syst.,
Fig. 13. (a) Original satellite image of a part of Greece. (b) Result after vol. 8, pp. 615–626, Oct. 2000.
adaptive Wiener filtering (best result with 5 2 5 support). (c) Result after [5] B. Reusch, M. Fathi, and L. Hildebrand, Soft Computing, Multi-
media and Image Processing—Proceedings of the World Automation
fuzzy filtering ( = 1, 5 iterations). (d) Result after fuzzy filtering ( = 3,
5 iterations). Congress. Albuquerque, NM: TSI Press, 1998, ch. Fuzzy Color
Processing for Quality Improvement, pp. 841–848.
[6] S. Bothorel, B. Bouchon, and S. Muller, “A fuzzy logic-based approach
for semiological analysis of microcalcification in mammographic im-
well. Only the fuzzy median (FM) gives a better MSE for . ages,” Int. J. Intell. Syst., vol. 12, pp. 819–843, 1997.
A closer inspection of Fig. 9 shows that the proposed filter better [7] F. Russo and G. Ramponi, “A fuzzy operator for the enhancement of
preserves details such as the grass (right side, just below the blurred and noisy images,” IEEE Trans. Image Processing, vol. 4, pp.
1169–1174, Aug. 1995.
building) and the background (left side, small buildings). Also [8] , “A fuzzy filter for images corrupted by impulse noise,” IEEE
the face is slightly sharper. The detail images in Fig. 10 con- Signal Processing Lett., vol. 3, pp. 168–170, June 1996.
firm these results. Note that the grass is better preserved by the [9] F. Russo, “Fire operators for image processing,” Fuzzy Sets Syst., vol.
103, no. 2, pp. 265–275, 1999.
proposed filter than using the fuzzy mean. The “boats” image [10] C.-S. Lee, Y.-H. Kuo, and P.-T. Yu, “Weighted fuzzy mean filters for
provides a different result. For low noise levels ( ), the pro- image processing,” Fuzzy Sets Syst., no. 89, pp. 157–180, 1997.
posed filter still performs best, but for higher noise levels, the [11] C.-S. Lee and Y.-H. Kuo, Fuzzy Techniques in Image Processing. New
York: Springer-Verlag, 2000, vol. 52, Studies in Fuzziness and Soft
AWFM2 filter gives the best results. Fig. 11 shows the filtered Computing, ch. Adaptive fuzzy filter and its application to image
images. The detail images of Fig. 12 reveal that the AWFM2 enhancement, pp. 172–193.
436 IEEE TRANSACTIONS ON FUZZY SYSTEMS, VOL. 11, NO. 4, AUGUST 2003

[12] F. Farbiz and M. B. Menhaj, Fuzzy Techniques in Image Pro- Etienne E. Kerre was born in Zele, Belgium, in
cessing. New York: Springer-Verlag, 2000, vol. 52, Studies in 1945. He received the M.Sc. and Ph.D. degrees
Fuzziness and Soft Computing, ch. A fuzzy logic control based in mathematics from Ghent University, Ghent,
approach for image filtering, pp. 194–221. Belgium, in 1967 and 1970, respectively.
[13] H. Haussecker and H. Tizhoosh, Handbook of Computer Vision and Ap- Since 1984, he has been a Lector and, since 1991,
plications. New York: Academic, 1999, vol. 2, ch. Fuzzy Image Pro- a Full Professor at Ghent University. In 1976, he
cessing, pp. 708–753. founded the Fuzziness and Uncertainty Research
[14] J. S. Lim, Two-Dimensional Signal and Image Processing. Upper Modeling Unit (FUM) and, since then, his research
Saddle River, NJ: Prentice-Hall, 1990, ch. Image Restoration, pp. has been focused on the modeling of fuzziness and
524–588. uncertainty, and has resulted in a great number of
[15] K. Arakawa, “Median filter based on fuzzy rules and its application to contributions in fuzzy set theory and its various
image restoration,” Fuzzy Sets Syst., pp. 3–13, 1996. generalizations, and in evidence theory. The theories of fuzzy relational
calculus and fuzzy mathematical structures owe a very great deal to him. Over
the years, he has also been a promoter of 16 Ph.D. degrees on fuzzy set theory.
His current research interests include fuzzy and intuitionistic fuzzy relations,
Dimitri Van De Ville (M’02) was born in Den- fuzzy topology, and fuzzy image processing. He has authored or coauthored
dermonde, Belgium, in 1975. He received the eleven books and more than 100 papers of his have appeared in international
Engineering and Ph.D. degrees in computer science refereed journals.
from Ghent University, Ghent, Belgium, in 1998 and Dr. Kerre is a referee for more than 30 international scientific journals, and is
2002, respectively. also Member of the Editorial Board of international journals and conferences
He worked in the Medical Image and Signal on fuzzy set theory. He was an Honorary Chairman at various international
Processing Group (MEDISIP) and the MultiMedia conferences.
Lab, both part of Department of Electronics and
Information Systems (ELIS), Ghent University. His
main research interests are in signal and image pro-
cessing, in particular, interpolation and resampling
related topics. Currently, he is working as a Senior Researcher at the Swiss
Federal Institute of Technology Lausanne (EPFL) in the Biomedical Imaging Wilfried Philips (S’90–M’93) was born in Aalst,
Group (BIG), Lausanne, Switzerland. Belgium, in 1966. He received the Diploma degree
in electrical engineering and the Ph.D. degree in
applied sciences from Ghent University, Ghent,
Mike Nachtegael was born in Sint-Niklaas, Belgium, in 1989 and 1993, respectively.
Belgium, in 1976. He received the M.Sc. degree From October 1989 to October 1998, he was
in mathematics from Ghent University, Ghent, with the Department of Electronics and Information
Belgium, in 1998. In the same year, he joined the Systems, the University of Ghent, for the Flemish
Fuzziness and Uncertainty Modeling Research Unit Fund for Scientific Research (FWO-Vlaanderen),
of Prof. E. Kerre, where he received the Ph.D. degree first as a Research Assistant and later as a Postdoc-
on fuzzy techniques in image processing in 2002. toral Research Fellow. Since November 1997, he has
Currently, he is active as a Postdoctoral Re- been a Lecturer with the Department of Telecommunications and Information
searcher in the Department of Applied Mathematics Processing, Ghent University. His main research interests are image and video
and Computer Science, Ghent University. After restoration, image analysis, lossless and lossy data compression of images and
secondary school, he published two reference video, and processing of multimedia data.
books on mathematics (1995) and on chemistry and physics (1996). He has
authored or coauthored more than 20 papers, he has coedited two books on
fuzzy techniques in image processing, he has coorganized three sessions at
international conferences and he was comanager of the International FLINS
2002 Conference.
Ignace Lemahieu (M’92–SM’00) was born in Bel-
gium in 1961. He graduated in physics and received
the Ph.D. degree in physics from Ghent University,
Dietrich Van der Weken was born in Beveren, Ghent, Belgium, in 1983 and 1988, respectively.
Belgium, in 1978. He received the M.Sc. degree He joined the Department of Electronics and In-
in mathematics from Ghent University, Ghent, formation Systems (ELIS), Ghent University in 1989
Belgium, in 2000. In September 2000, he joined the as a Research Associate with the Fund for Scientific
Department of Applied Mathematics and Computer Research (F.W.O.-Flanders), Belgium. He is now a
Science, Ghent University, where he is a member of Professor of Medical Image and Signal Processing
the Fuzziness and Uncertainty Modeling Research and Head of the MEDISIP Research Group. His re-
Unit working toward the Ph.D. degree with a thesis search interests comprise all aspects of image pro-
on fuzzy techniques in image processing under the cessing and biomedical signal processing, including image reconstruction from
promotorship of Prof. E. Kerre. projections, pattern recognition, image fusion, and compression. He is the coau-
One of his main research topics is the measure- thor of more than 200 papers.
ments of similarity between images. He has authored or coauthored 14 papers, Dr. Lemahieu is a Member of SPIE, the European Society for Engineering
he has co-edited one book on fuzzy techniques in image processing, and orga- and Medicine (ESEM), and the European Association of Nuclear Medicine
nized one session at an international conference. (EANM).