Signal Processing
Members:
Source Separation
Researchers:
O. Deniz Gençağa, O. Gökhan Sezer, Ahmed Serdaroğlu, Nazlı Güney, Kerem Harmancı,
Ercan Kuruoğlu, Ayşın Ertüzün
Description:
Source separation problem is the process
of separating background sources from the additive or convolutive mixtures of these sources, i.e. the observations.
The research activities of our group have concentrated on source separation
methods where
independent component analysis and Bayesian techniques are used.
Our objectives are echo cancellation, astrophysical source separation, texture
defect detection, and recognition of planar objects.
Publications:
A. Serdaroğlu, A. Ertüzün and A. Ercil, Defect Detection in Textile Fabric Images
Using Wavelet Transforms And Independent Component Analysis accepted, Pattern Recognition and Image Understanding : New Technologies, PRIA-7-2004, Oct. 18-23, 2004,
O.G. Sezer, A. Ertüzün and A. Ercil,
Independent Component Analysis for Texture Defect Detection, Proceedings of
the 6th German-Russian Workshop "Pattern Recognition and Image
Understanding" (OGRW-6-2003), Katun Village,
-Novosibirsk, Altai region, Russian Federation, August, 25-30, 2003, pp.
210-213.
O.G. Sezer, A. Ertüzün and A. Ercil, Textile Fabric Defect Detection using
Independent Component Analysis, Submitted to Image and Vision Computing, 2003.
Adaptive Filters
Adaptive filters constitute an important part
of statistical signal processing. When signals to be processed are evolved from
unknown statistics, adaptive filters can be suitably used. The research
activities of our group have concentrated on development of adaptive algorithms
and structures as well as application of adaptive algorithms and adaptive
systems to image and signal processing problems in various fields.
Researchers: Korkut Uluaydın, Idil Haskan, O. Deniz Gençağa, Ayşın Ertüzün
Description: Acoustic echo is generated when the
microphone and the loudspeaker are placed close to each other in an enclosure.
This is an important problem which deteriorates the quality of communication
especially in hands-free telephone applications. Our goal is to eliminate or
reduce the echo and enhance the quality of conversation to achieve the
standards given ITU-T:G.167.
Adaptive 2-D Lattice Filters and Applications
to Image Processing
Researchers: C.Mehmet Hendekli, Ayşın
Ertüzün
Description: A new 2-D lattice
filter structure in the so-called delta domain, which avoids possible numerical
instabilities that may occur in traditional q-domain
based algorithms is developed and used to enhance images both single and multichannel cases.
Publications:
İ. Haskan
and A. Ertüzün, System Identification Using
Orthogonal Functions And Application To Acoustic Echo
Cancellation, in Proc. Int. Conf. On Acoustic, Speech and
Signal Processing (ICASSP1999),
C. M. Hendekli
and A. Ertüzün, 2-D Delta Domain Lattice Filter and
Its Application to Adaptive Image Enhancement, IEEE Trans. Signal Proc. Vol 47, No1. ,
Jan 1999, pp.266-273.
C.
M. Hendekli and A. Ertüzün,
Design of a Multichannel Two-Dimensional
Delta-Domain Lattice Filter For Noise Removal, IEEE Trans. on Signal Processing, Vol. 49, No.7, July 2001,
pp.1581-1593.