Analysis and Interpretation of Signals

Garant předmětu:

Lecturer:

Tutor:

Supervisor:

Department of Biomedical Informatics

Synopsis:

The subject covers special methods of signal processing, mainly processing 1D

and ND signals (images), interpretation of measured data, sampling, Fourier and

wavelet transformation, filtration in time and frequency domain. Image

processing represents significant part of the subject: image preprocessing,

histograms, histogram equalisation, edge detection, segmentation, morphology,

image compression. Further, utilizing neural networks for signal processing is

discussed.

Requirements:

Exam: Elaboration of a real task which is solved within master study project or

diploma thesis using method of the subject. The task is presented in the form of

web page.

Syllabus of lectures:

- Analysis and interpretation of signals. Basic terms: signal, analysis,

synthesis, interpretation

- Recognition and classification.Classifier and its setting, feature and

structural methods of recognition, evaluating of classification quality

- Feature methods for reconition. classification with minimal error,

classification with minimal distance, disciminant function

- Parametric and non-parametric methods of estimations, cluster analysis,

hierarchical and nonhierarchical methods.

- Structural methods, structural description, selection of primitivs and

relations, description of time-domain signals, description of images

- Formal languages, gramatics, automatons

- Syntactic analysis, inference, utilizing gramatics for recognition, utilizing

semantic information, deformation schema for classification

- Processing and recognition of images, computer vision, image preprocessing,

contrast transformation, geometric corections and transformations, filtration,

gradient operators

- Image segmentation, object description, area and edge representation, object's

shape features, texture, analysis of 3D scene, scene interpretation

- Recogniction of vocal signals, information and phonetics aspects, acustic

analysis and feature selection, recognition of isolated words and fluent speech

- Adaptive and learning algorithms, artificial neural networks, principles of

learning, memory, biological motivation of ANN, paradigms of ANN {MLP, RBF,

Hopfield, ...)

- Evolutionary techniques, genetic algorithms, genetic programming

- Prediction of time-domain signals, feature selection, preparation of sets for

training, utilizing ANN

- Knowledge and expert systems, data mining, diagnostic expet systems

Syllabus of tutorials:

Study Objective:

The subject covers special methods of signal processing, mainly processing 1D

and ND signals (images), interpretation of measured data, sampling, Fourier and

wavelet transformation, filtration in time and frequency domain. Image

processing represents significant part of the subject: image preprocessing,

histograms, histogram equalisation, edge detection, segmentation, morphology,

image compression. Further, utilizing neural networks for signal processing is

discussed.

Study materials:

[1]Sonka M., Fitzpatrick J. M.: Medical Image Processing and Analysis, SPIE -

The International Society for Optical Engineering, Bellingham, WA, USA, 2000

[2]Sonka M., Hlaváč V., Boyle R. D.: Image Processing, Analysis and Machine

Vision, Boston, USA, 1998

[3]Prince J. L., Links J. M.: Medical Imaging Signals and Systems, Pearson

Education, NJ, USA, 2006

Note:

Further information:

Course may be repeated

No time-table has been prepared for this course

The course is a part of the following study plans: