Research Activities
My research is devoted
to biomedical image enhancement, which falls within the generic multidisciplinary
area of information engineering, known as Digital Signal and Image Processing (DSP).
There are many applications in which signals are converted into digital form
and then digital signal processing methods are applied. In the case of digital
image processing, the digital signal is two-dimensional. My work presents some of the
basic tools for digital image enhancement: image denoising, reconstruction of missing
or corrupted parts of images and image resolution enhancement.
Denoising is the first important part of image
enhancement. The procedures for image denoising involve wavelet decomposition,
thresholding and wavelet reconstruction. This approach is known as denoising
using wavelet transforms. Nowadays the wavelet transform is a fundamental tool,
used in a wide range of digital signal processing techniques, including image
compression.
The next step in digital image enhancement is
the reconstruction of degraded parts of the image. This is closely related to
image segmentation, boundary detection, signal modelling and signal prediction.
The next part of my research is devoted to methods for image component
reconstruction. There are linear, nonlinear, and probabilistic algorithms
described in my papers (below). The linear algorithms are based on autoregressive
modelling and bilinear interpolation. Nonlinear methods include image subregion
feature extraction and classification. Signals containing more random components can be more completely
described by their probability distributions. Therefore Bayesian probabilistic
methods are important in the analysis of two-dimensional signals. The Bayesian
approach has been adopted and used for computing the maximum a posteriori
estimate of the original image given the degraded image. This method forms an alternative
approach to deterministic models, which are useful for good predictable
signals.
Finally, a fundamental problem encountered in
the digital processing of both one-dimensional and two-dimensional signals is
the selection of the signal resolution. This defines the sampling period in the
case of time series or the pixel size in the case of images. Changing the
resolution of a signal or image allows both global and detailed views of
specific one-dimensional or two-dimensional signal components. Signal and image
resolution enhancement is therefore also a fundamental problem in signal
analysis.
There are numerous existing models and
algorithms for digital image enhancement and this field of information engineering is currently
very active. My papers describe known linear methods and newly
designed methods adapted to the specific properties of biomedical images, which
are the main application in my work.
Publications
[1] Ptáček J., Stříbrský J., Procházka A.: Wavelet Transform in Signal Processing,
the 4th international
conference on Process Control 2000, 11-14 June, 2000, Kouty nad Desnou, Czech Republic (in English)
[1] Ptáček J., Slavík M.: Wavelet Transform, seminar at the Department of Mathematics, ICT Prague, 11th December, 2000 (in Czech)
[2] Ptáček J.: Image Components Analysis and Reconstruction, seminar at the Department of Computing and Control
Engineering, ICT Prague, 1st June, 2001 (in Czech)
[3] Ptáček J.: Enhancement of Biomedical Images, seminar at Brunel University London, Department of
Electronics and Computer Engineering, 8th April, 2002, London (in English)
[4] Ptáček J., Procházka A.: Image Resolution Enhancement, seminar at Brunel University London, Department
of Electronics and Computer Engineering, 17th June, 2002, London (in English)
[5] Ptáček J.: Bayesian Methods and Wavelet Transform in Image Components Reconstruction, seminar at
Brunel University London, Department of Electronics and Computer Engineering, 12th August, 2002, London (in English)
[6] Ptáček J.: Digital Image Enhancement in Biomedical Applications, conference of PhD students at the
Faculty of Chemical Engineering, ICT Prague, 14th November, 2002, Prague (in English)
[7] Ptáček J.: Digital Image Enhancement Using Wavelet Decomposition, seminar at the Department of Computing and Control
Engineering, ICT Prague, 23rd May, 2003 (in Czech)
[8] Ptáček J., Procházka A.: Digital Image Processing Using Wavelet Transform, seminar at Brunel University London, Department
of Electronics and Computer Engineering, 5th September, 2003, London (in English)
Science Reports
[1] Ptáček J.: Review of Jiří Dluhoš's master thesis, topic: "Kinetic Data Processing Using Artificial Neural Networks", for the Department
of Organic Technology, ICT Prague, 30th May 2000 (in Czech)
[2] Procházka A., Mudrová M., Kukal J., Bártová D., Stříbrský J., Kolínová M., Pánek M., Ptáček J.: Analysis and Evaluation of Students'
Knowledge and Teaching of Information Technology at ICT Prague, report for the internal grant of ICT Prague 445 01 0015,
29th January 2001 (in Czech)
[3] Ptáček J.: Review of Martin Novotný's master thesis, topic: "Digital Signal Processing Algorithms
Influence on Measurement Indeterminateness of Selected Signal Parameters, 5th February 2003, for the Department of Measurements,
CTU Prague (in Czech)