![](bg6.png)
5 DISCUSSION AND
CONCLUSIONS
In this paper we have presented a novel unified frame-
work under which the histogram based texture de-
scription methods such as local binary pattern and
MR8 descriptors can be explained and analyzed. This
framework allows for systematic comparison of dif-
ferent texture descriptors and the parts that the de-
scriptors are built of. Such novel approach can be
useful in analyzing texture descriptors since they are
usually presented as a sequence of steps whose rela-
tion to other texture description methods is unclear.
The framework presented in this work allows for ex-
plicitly illustrating the connection between the parts
of the LBP and MR8 descriptors and experimenting
with the performance of each part.
The filter sets and vector quantization techniques
for LBP, MR8 and Gabor filter based texture descrip-
tors were compared in the this paper. In this com-
parison it was found out that the local derivative fil-
ter responses are both fastest to compute and most
descriptive. This somewhat surprising result further
attests the previous findings that texture descriptors
relying on small-scale pixel relations yield compara-
ble or even superior results to those based on filters of
larger spatial support (Ojala et al., 2002), (Varma and
Zisserman, 2003).
When comparing the different vector quantization
methods, codebook based quantization was discov-
ered to be slightly more descriptive than thresholding
in most cases. Finally, the preliminary experiments on
combining local derivative and Gabor filter responses
showed that these filter sets may be complementary
and may yield better performance than either of the
sets alone.
Not only does the presented framework contribute
to understanding and comparison of existing texture
descriptors but it can be utilized for more systematic
development of new, even better performing meth-
ods. The framework is simple to implement and to-
gether with the publicly available KTH-TIPS2 image
database it can be easily used for comparing novel
descriptors with the current state-of-the-art methods.
We believe that further advances in both the filter bank
and vector quantizer design are possible, especially as
new invariance properties of the descriptors are aimed
for.
REFERENCES
Caputo, B., Hayman, E., and Mallikarjuna, P. (2005). Class-
specific material categorisation. In Proc. ICCV 2005,
pages 1597–1604.
Clausi, D. A. and Jernigan, M. E. (2000). Designing Gabor
filters for optimal texture separability. Pattern Recog-
nition, 33(11):1835–1849.
Gray, R. M. and Neuhoff, D. L. (1998). Quantization. IEEE
Trans. Image Process, 44(6):2325–2383.
Grigorescu, S. E., Petkov, N., and Kruizinga, P. (2002).
Comparison of texture features based on Gabor filters.
IEEE Trans. Image Process., 11(10):1160–1167.
Leung, T. and Malik, J. (2001). Representing and recog-
nizing the visual appearance of materials using three-
dimensional textons. Int. J. Comput. Vision, 43(1):29–
44.
Mallikarjuna, P., Fritz, M., Targhi, A. T., Hayman,
E., Caputo, B., and Eklundh, J.-O. (2006).
The KTH-TIPS and KTH-TIPS2 databases.
http://www.nada.kth.se/cvap/databases/kth-tips/.
Manjunath, B. S. and Ma, W.-Y. (1996). Texture features
for browsing and retrieval of image data. IEEE Trans.
Pattern Anal. Mach. Intell., 18(8):837–842.
Ojala, T., Pietik¨ainen, M., and M¨aenp¨a¨a, T. (2002). Mul-
tiresolution gray-scale and rotation invariant texture
classification with local binary patterns. IEEE Trans.
Pattern Anal. Mach. Intell., 24(7):971–987.
Ojala, T., Valkealahti, K., Oja, E., and Pietik¨ainen, M.
(2001). Texture discrimination with multidimensional
distributions of signed gray-level differences. Pattern
Recognition, 34(3):727–739.
Pudil, P., Novovicov´a, J., and Kittler, J. (1994). Floating
search methods in feature selection. Pattern Recogni-
tion Letters, 15(10):1119–1125.
The Local Binary Pattern Bibliography (2007).
http://www.ee.oulu.fi/mvg/page/lbp bibliography.
Tuceryan, M. and Jain, A. K. (1998). Texture analysis. In
Chen, C. H., Pau, L. F., and Wang, P. S. P., editors,
The Handbook of Pattern Recognition and Computer
Vision (2nd Edition), pages 207–248. World Scientific
Publishing Co.
Varma, M. and Zisserman, A. (2003). Texture classification:
Are filter banks necessary? In Proc. CVPR 2003, vol-
ume 2, pages 691–698.
Varma, M. and Zisserman, A. (2004). Unifying statistical
texture classification frameworks. Image Vision Com-
put., 22(14):1175–1183.
Varma, M. and Zisserman, A. (2005). A statistical approach
to texture classification from single images. Interna-
tional Journal of Computer Vision, 62(1–2):61–81.
Yan, S., Wang, H., Tang, X., and Huang, T. S. (2007). Ex-
ploring feature descriptors for face recognition. In
Proc. ICASSP 2007.
VISAPP 2008 - International Conference on Computer Vision Theory and Applications
512