Course detail
Advanced Signal Processing
FIT-MZSAcad. year: 2013/2014
Formalised inverse filtering and restoration of signals. Wiener filter, constrained deconvolution and further higher restoration approaches. Kalman filtering, scalar and vector formulation, system modelling based on Kalman filtering. Adaptive filtering and identification, algorithms of adaptive filters, typical applications of adaptive filtering. Multirate systems. Non-linear filtering: polynomial filters, rank filters, homomorphic filtering and deconvolution, non-linear matched filters. Signal processing by neural networks. Time-frequency analysis, wavelet transform and its applications. Concept of multidimensional signal and spectrum, 2D and 3D Fourier transform, discrete unitary multidimensional transforms. Applications in formalised image processing: restoration approaches, tomographic reconstruction from projections, 3D reconstruction from stereo data.
Language of instruction
Number of ECTS credits
Mode of study
Guarantor
Learning outcomes of the course unit
Prerequisites
Co-requisites
Planned learning activities and teaching methods
Assesment methods and criteria linked to learning outcomes
Course curriculum
- Syllabus of lectures:
- Formalised inverse filtering and restoration of signals. Wiener filter: classical and discrete formulation
- Constrained deconvolution, deconvolution with impulse response optimisation, maximum posterior-probability method
- Kalman filtering, scalar and vector formulation, system modelling based on Kalman filtering
- Concept of adaptive filtering and identification, algorithms of adaptive filtering
- Typical applications of adaptive filtering: system identification and modelling, linear adaptive prediction, adaptive noise and interference suppression
- Multirate systems of digital signal processing, multirate filter banks
- Non-linear filtering: polynomial filters, rank filters, homomorphic filtering and deconvolution, non-linear matched filters
- Signal processing by neural networks: learning neuronal filters and classifiers, restoration by feed-back neuronal networks
- Time-frequency analysis, wavelet transform and its applications in processing and compression of signals
- Concept of multidimensional signal and spectrum, 2D and 3D Fourier transform. Discrete unitary 2D transforms: cosine, Hadamard and Walsh, Haar and 2D wavelet tr.
- Applications of signal-theory-based approaches to formalised image processing: restitution and restoration approaches, formalised image segmentation
- Tomographic methods of image reconstruction from 1D projections
- Motion analysis and 3D reconstruction from stereo data
- Simulation of discrete Wiener filter and evaluation of efficiency in stationary case
- Simulation of a 3rd order Kalman filter and comparison with the above Wiener filter in stationary environment
- Simulation of adaptive filters of RLS and LMS type as applied to a system modelling. Comparison of both results in stationary and slowly varying environment
- Wavelet transform: application to analysis and denoising of a signal, verification of compression ability
- Restoration of blurred and noisy image by pseudoinversion and by 2D classical Wiener filter - comparison of results
- 2D image reconstruction from tomographic data (1D projections) via frequency domain - evaluation of artefacts
- Learning 2D neuronal filter: applied for texture analysis. Comparison with feature-oriented classification
Syllabus of computer exercises:
Work placements
Aims
Specification of controlled education, way of implementation and compensation for absences
Recommended optional programme components
Prerequisites and corequisites
Basic literature
Recommended reading
Classification of course in study plans
- Programme IT-MSC-2 Master's
branch MBI , 0 year of study, summer semester, elective
branch MBS , 0 year of study, summer semester, elective
branch MGM , 0 year of study, summer semester, elective
branch MIN , 0 year of study, summer semester, elective
branch MIS , 0 year of study, summer semester, elective
branch MMI , 0 year of study, summer semester, elective
branch MMM , 0 year of study, summer semester, elective
branch MPV , 0 year of study, summer semester, elective
branch MSK , 0 year of study, summer semester, elective
Type of course unit
Lecture
Teacher / Lecturer
Syllabus
- Formalised inverse filtering and restoration of signals. Wiener filter: classical and discrete formulation
- Constrained deconvolution, deconvolution with impulse response optimisation, maximum posterior-probability method
- Kalman filtering, scalar and vector formulation, system modelling based on Kalman filtering
- Concept of adaptive filtering and identification, algorithms of adaptive filtering
- Typical applications of adaptive filtering: system identification and modelling, linear adaptive prediction, adaptive noise and interference suppression
- Multirate systems of digital signal processing, multirate filter banks
- Non-linear filtering: polynomial filters, rank filters, homomorphic filtering and deconvolution, non-linear matched filters
- Signal processing by neural networks: learning neuronal filters and classifiers, restoration by feed-back neuronal networks
- Time-frequency analysis, wavelet transform and its applications in processing and compression of signals
- Concept of multidimensional signal and spectrum, 2D and 3D Fourier transform. Discrete unitary 2D transforms: cosine, Hadamard and Walsh, Haar and 2D wavelet tr.
- Applications of signal-theory-based approaches to formalised image processing: restitution and restoration approaches, formalised image segmentation
- Tomographic methods of image reconstruction from 1D projections
- Motion analysis and 3D reconstruction from stereo data
Exercise in computer lab
Teacher / Lecturer
Syllabus
- Simulation of discrete Wiener filter and evaluation of efficiency in stationary case
- Simulation of a 3rd order Kalman filter and comparison with the above Wiener filter in stationary environment
- Simulation of adaptive filters of RLS and LMS type as applied to a system modelling. Comparison of both results in stationary and slowly varying environment
- Wavelet transform: application to analysis and denoising of a signal, verification of compression ability
- Restoration of blurred and noisy image by pseudoinversion and by 2D classical Wiener filter - comparison of results
- 2D image reconstruction from tomographic data (1D projections) via frequency domain - evaluation of artefacts
- Learning 2D neuronal filter: applied for texture analysis. Comparison with feature-oriented classification