Faculty of Electrical Engineering and Information Technologies Skopje, Macedonia

in partnership with

Faculty of Electrical Engineering and Computer Science Maribor, Slovenia and

Faculty of Electronics and Telecommunications Poznan, Poland

 

Structure


Core courses:

Elective courses:

Statistical signal processing


Course Objective

The goal of this course is to provide a comprehensive coverage of statistical signal processing methods and tools, including the algorithms for various applications.

 

Course prerequisite:

This course requires basic knowledge of probability.

 

Texts:

[1]        M. Hayes, “Statistical Digital Signal Processing and Modeling”, Wiley, 1996.

[2]        D. G. Manolakis, V. K. Ingle, S. M. Kogon, “Statistical and Adaptive Signal Processing: Spectral Estimation, Signal Modeling, Adaptive Filtering and Array Processing”, Artech House, 2005.

 

Course Outline: (Outline syllabus)

Random vectors: definition, moments, characteristic functions, multidimensional Gaussian distribution. Discrete time random processes: definition, stationarity and ergodicity, autocorrelation and power spectral density, Markov and Gauss-Markov processes. Estimation of unknown parameters: MVUE, ML, LS. Estimation of random parameters: MAP, MMSE, orthogonality principle. Optimal estimation of discrete time random processes: Wiener and Kalman filters. Parametric models of discrete time random processes: AR, MA, ARMA models. Spectral analysis of discrete time random processes: periodogram, correlogram, methods based on parametric models, high resolution methods. Adaptive signal processing: steepest descent, LMS, RLS algorithms. Array signal processing: beam forming, high resolution methods. Applications of statistical signal processing methods and algorithms.

 

Learning Outcomes

Upon completion of this course, students will be able to:

-understand and know how to implement the methods and algorithms of statistical signal processing: parameter estimation, random parameter estimation and estimation of random processes, and adaptive signal processing 

-identify the engineering problems that can be put into the frame of statistical signal processing, 

-solve the identified problems using the techniques learned through this course, and

-apply the fundamental ideas of statistical signal processing to study further and make significant contributions to the theory and the practice of statistical signal processing.   

 

Course methodology:

The course concepts will be taught by the instructor. Students will have to prepare and teach certain lectures in class. A research project will be assigned to each student/group of students. At the end of the semester, each student/group will present their project. There will be a final exam covering the principal methods and algorithms learned in the course.

 

Grading Method:

The grading will be based on the exam and research project. Specifically:

Exam: 50%

In/class presentation of lecture: 10%

Project: 40%

top ↑

DSP in transform domain


Course Aims 

Digital Signal Processing (DSP) is a fascinating blend of mathematics, computation and practical implementation. It is the bedrock upon which all modern digital communication systems are built.

This course mainly  focuses on the basic concepts and techniques for processing one and two dimensional signals in transform domain and includes in-depth treatment of different discrete transforms and various  fast algorithms giving the necessary mathematical background for practical implementations of the theoretical concepts . The final part of the course examines several typical transform domain application as filtering, spectral estimation, coding, adaptive filtering, multirate signal processing.

By the end of the course, students will be familiar with the most important methods and algorithms  in DSP transform-domain processing.

 

Course prerequisite:

Basic knowledge of signals, systems and signal processing.

 

Texts:

[1]      D. F. Elliot and K. Ramamohan Rao, Fast Transforms: Algorithms, Analyses, Applications, Academic Press. Orlando FL, 1982.

[2]      V. K. Madisetti and D. B. Williams, Digital Signal Processing, CRC Press, 1998. 

[3]      H. S. Malvar, Signal Processing with Lapped Transform, Artech House, Boston MA, 1. 

 

Course Outline: (Outline syllabus)

Brief revue of signal models. Definitions and basic properties of discrete transforms. Orthogonal discrete transforms: Fourier(DFT), Hartley(DHT), Karhunen-Loeve (KLT), Cosine (DCT), Lapped (LOT), wavelet(WLT), Walsh-Hadamard (WHT). Two- dimensional transforms. Fast algorithms: concept and selected examples. Applications in signal processing: filtering, spectral estimation, coding, adaptive filtering, multirate signal processing.

 

Learning Outcomes

A student who has met the objectives of the course will be able to:

·   The students should be able to achieve the above goals in a group effort while maintaining individual accountability

·   The student should be able to communicate his results in a clear and precise manner

 

Course methodology:

The course concepts will be taught by the instructor. A research project will be assigned to group of students. They  will be expected to do a bit of independent reading, a presentation in class, and a short report. There will be a final exam covering the principal methods and algorithms learned in the course.

 

Grading Method:

The grade will be determined from class attendance,  homeworks, and a final project. Homeworks will be issued periodically (5 or 6 during the semester). The students  are encouraged to collaborate on them.

top ↑

Filter banks and wavelets


Various transform-domain techniques are proposed in order to analyze the signals with aim to identify their nature and most important characteristics. Probably the most famous and most used is Fourier transform. However in some application the use of the Fourier transform entails a difficulty, namely any time-local information about the signal is lost. The wavelet transform, its natural framework of multiresolution analysis and its discrete-time counterpart (the discrete wavelet transform performed by multirate filter banks) allow for better analysis of signals, by looking at them at various scales or resolutions. As results of this property, there is increasing number of applications where wavelet transform is used. For instance, in image coding, wavelet based coding methods already outperform current JPEG standards.

This course will provide both a mathematical background in wavelets and an introduction to their discrete-time implementation (multirate filter banks).

 

Pre-requisite:

Digital Signal Processing course

 

Literature:

[1]     G. Strang and T. Nguyen, Wavelets and filter banks,   Wellesley-Cambridge Press

[2]     P.P. Vaidyanathan, Multirate Systems and Filters Banks, Prentice Hall, 1993

[3]     M. Vetterli and J. Kovacevic, Wavelets and Subband coding, Prentice Hall, 1995

 

Outline syllabus

  • Introduction to the fundamental theory of multirate signal processing: decimation, interpolation, and sampling rate conversion

  • Two-channel filter banks: quadrature-mirror filter banks (QMF), perfectly reconstructing, paraunitary, biorthogonal and linear phase filter banks,

  • M-channel filter banks used as subband coding or transmultiplexing filter banks.

  • Polyphase structures for two-channel and M-channel filter banks.

  • Lattice structures for Linear Phase PR QMF Banks

  • Discrete Wavelet Transform (DWT) and its relations to multirate filter banks. The Short-Time Fourier Transform. The Wavelet Transform. Discrete-Time Orthonormal Wavelets. Continuous-Time Orthonormal Wavelet Bases.

  • Lifting implementations of wavelet transform

  • Applications of wavelets for signal analysis and compression.

 
Learning Outcomes

Having successfully completed this course, student will be able to: 

  • Understand the basics of multirate signal processing

  • understand the idea of a different types of filter banks

  • gain an overview of their design methods

  • understand how to design perfect reconstruction filter banks

  • understand the fundamentals of wavelet theory

  • become familiar with the most commonly used wavelets

  • understand the link between design of filter banks and construction of discrete and continuous-time bases for efficient signal analysis.

  • analyse and compress a signal using wavelets.

  • work in team

  • research, present, and report a selected project within a specified time.

  • think critically, ask questions, and apply problem-solving techniques.

 

Course methodology:

Collaborative learning model will be implemented in this course. Accent will be given to the team work. For efficient implementation of the proposed collaborative model it is essential that student come to class familiar with the material that will be discussed that day. The class is split to short lectures  for teaching and for group work of the student. The lectures will focus on the most important or most difficult concepts in the reading. Students will spend most of their class time working in groups on problems. 

After each class homework will be assigned. In most of the homeworks the concepts discussed in the class should be implemented in Matlab. One project will be assigned to each group of students. The project will be presented by the group members of the end of the semester.

 

Grading Method:

Main course activities will be evaluated and grading will be as follows:

Exam: 60 %

Homework: 20 %

Project: 20 %

top ↑

Real time DSP lab


Course Aims  

With the widespread usage of DSPs as part of real-time embedded systems, the breadth of applications in which DSPs can be used is large. These applications include, but are not limited to, modems, faxes, data transmission, data encryption, speech and image processing and compression, vehicle navigation, automotive control, seismic and spectral analysis, radar and sonar, ECG monitoring, digital audio and music, hearing aids, digital cellular phones, and video telephones.

 

The objectives of the proposed course are to provide the students with knowledge and hands-on experience in translating DSP concepts into real-time software for embedded systems using DSP boards. The proposed course emphasizes designing and implementing real-time software for embedded systems through the programming of a special type of micro-processor, the DSP, or Digital Signal Processor. Laboratory exercises will be based on writing C and assembly language software for selected DSP boards that are used in current consumer products, and interfacing the DSPs to external devices for test and measurement.

 

Course prerequisite:

This course requires basic knowledge of: C programming, Signals and systems

 

Texts:

[1]  Sen M. Kuo, Bob H. Lee, Wenshun Tian: "Real-Time Digital Signal Processing: Implementations and Applications", Wiley, Prentice Hall

[2]  Sen M_Kuo,Woon-Seng S_Gan: "Digital Signal Processors - Architectures, Implementations, and Applications", (Amazon)

[3]  DSP manufacturer manuals

 

Course Outline: (Outline syllabus)

Introduction to digital signal processors, general characteristics and real-time applications, DSP families, general rules for selecting a DSP.

Detailed architecture of a selected DSP, instruction set, development environment.

Integer and floating point arithmetic overview with emphasis on DSP arithmetic characteristics.

Overview of DSP peripherals.

Application examples, libraries, real-time operation optimization.

DSP/BIOS RT Kernel

 

Learning Outcomes

Having completed this unit, students will be able to demonstrate:

  • Knowledge of a digital signal processor architecture and it’s influence on real-time digital signal processing

  • Experience using equipment and tools commonly used in industry and/or experimental laboratory

  • Understanding of the process of converting ideas and algorithms into real-time working hardware with real-time signals

 Also, upon completion of this course, student should be able to:

  • work in team and individually

  • prepare for, perform and report experimental work

  • self-organize to acquire knowledge, implement it and track the progress through a real-life project(s)

 Course methodology:

The course will be given by lectures, readings, discussions, laboratory experiments and mini-project(s). After the general introductory lectures given by the instructor, students will be studying chapters from the text-book(s), datasheet and manuals and preparing written work plans for performing the assigned experiments. At the same time they will be exposed to certain examples and problems where real-time signal processing is needed to choose from for their project assignment. The work plans will consist of homework assignment as theoretical preparation and practical work preparation, and will be discussed and graded in class prior lab. experiment. After the lab. each student will prepare a report. After achieving the basic level of mastering the required tools and techniques, students will start working on their projects and will be giving regular reports on their progress that will be discussed and graded in class. The course will end with a final report, presentation and demonstration of their work.

 

Grading Method:

The grading will be based on student’s performance in written work plans and discussions, laboratory work, reports preparation and discussion and project report presentation and demonstration. Specifically:

Work plans: 20%

Reports: 20%

Laboratory: 20%

Project: 40%

top ↑

Biomedical signal analysis and processing


Course Description:

Nowadays, signal processing come in widely in all areas of technical and technological development. Parallel with the new achievements in signal processing, a row of disciplines develop too. So, on the demand for developing new techniques adequate to the signal processing characteristic is paid more and more attention.

Biomedical signal analysis and processing is focused on acquisition and processing of signals that come from living systems, and allow us to find out about their state, therefore its accurate interpretation has significant value for clinicians and researches.

The aim of this course is to demonstrate the biomedical signals, their origins and sources, use of techniques in processing, according to their characteristics, as noise reduction techniques spectral analysis and filtering, digital biomedical signal acquisition and processing, their compression, use of wavelet analysis. Special attention is focused on characteristics and analysis of ECG, EEG and EMG signals.

 

Course prerequisite:

Basic knowledge of signal processing is required for the course.

 

Texts:

[1]   K. Najarian, R. Splinter, “ Medical Signal and Image Processing”, CRC Press, 2006

[2]   J. D. Enderle, S. M. Blanchard, J. D. Bronzino, „Introduction to Biomedical Engineering“, Elsevier Press

[3]   J. D. Bronzino, editor, “Biomedical Engineering Fundamentals”, CRC Press, 2006

 

Course Outline: (Outline syllabus)

Signals, introduction, definitions

Use of techniques in processing – Fourier transform, wavelet transform, other methods

Clustering and classification of signals

Modulation and demodulation

Signal acquisition

Noise reduction techniques, time averaging, spectral analysis and filtering, optimal filtering

Biomedical signals, origins, sources, and properties

Characteristics of cell and tissue from electric aspect

Analysis of biosignals with low level in presence of noise

Digital biomedical signal acquisition and processing

Compression of digital biomedical signals

Wavelet analysis in  digital biomedical signals

Characteristics and analysis of ECG, EEG and EMG signals

Other biomedical signals

 

Learning Outcomes

With successfully completed a learning process, the student will be able to demonstrate knowledge and understanding of:

·     Signal definition and understanding of their characteristics

·     Techniques for signal processing

·     Their analysis

·     The characteristics of biomedical signals

·     Use of techniques for analysis and processing of biomedical signals according their characteristics

·     Special overview and analysis of ECG, EEG and EMG signals

·     Another biomedical signals

Also, the student should be capable to:         

·     Identify and formulate different problems connected with the material

·     Use the theoretical knowledge and techniques during research

·     Analyze critically the results and bring appropriate conclusions

·     Research, present, and report a selected project within a specified time

·     Work either individually or in team

 

Course methodology:

The course methodology contains the ways of the presentations of material. It will be combined from lectures and paper reading. The lectures are taught by the instructor, but some of them can be given by the students too. In the same time, students should read articles, it is necessary each of the student to have read several papers. So, organized in teams, they can select adequate project topics. During the semester, several seminars will be organized, so the students can present read articles in front of the class, and their current work on their projects. The students are expected to take active part either in the lecture or paper discussions. All the projects are presented at the end of semester by the student or team.

 

Grading Method:

The grading method will be according to the student's activities in class participation, lecture and paper presentation, and the making and presentation of research project. Specifically:

Class participation:                         10%

Lectures and paper presentation:     30%

Project:                                         60%.

top ↑

Biomedical image processing


Course Description

This course will offer the students the basics of biomedical image processing through the coverage of the physical principles of biomedical imaging and the image processing basics and techniques for image enhancement, compression, segmentation, registration and motion analysis. Students will learn the fundamentals behind image processing methods and algorithms with an emphasis on biomedical applications.

 

Course Prerequisite:

This course requires basic knowledge of signal processing.

 

Texts:

[1]     Kayvan Najarian, Robert Splinter, Biomedical Signal and Image Processing, CRC Press, 2006

[2]     Rafael C. Gonzalez, Richard E. Woods, Digital Image Processing, Prentice Hall, 3rd ed., 2007

[3]     Selected articles will be put on reserve

 

Course Outline: (Outline Syllabus)

 

  • Basic concepts in digital image processing: Image representation in the spatial and frequency domains; Digitization; Visual perception; Noise and image quality; Components of an image processing system

  • Biomedical imaging modalities: physics of X-rays, the Fourier slice theorem, x-ray and CAT imaging; Physical and physiological principles of  magnetic resonance and MR imaging; Physical and physiological principles of  ultrasound and ultrasound imaging

  • Basics of morphological image processing: Introduction; Logic operations on digital images; Dilation and erosion; Openning and closing

  • Filtering, enhancement and restoration: Intensity modifications; Mask processing; Spatial frequency processing; Wavelet denoising

  • Motion: Introduction; Motion quantification from image sequences; Application for measuring dynamic biological phenomena

  • Basics of image compression: Introduction; Lossless compression; Lossy compression

  • Edge detection and image segmentation: Edge detection; Thresholding; Region-based segmentation; Segmentation by morphological watersheds; Application of motion for segmentation

  • Rigid and non-rigid image registration: Introduction and transformations; Match metrices; Optimization and interpolation; Robustness

 

Learning Outcomes

Having successfully completed this course, you will be able to:

  • explain the fundamentals of: image formation and acquisition, image representation in the spatial and frequency domains, the respective roles of sampling, quantization, transformation and the HVS

  • explain the basic mechanisms and modalities of biomedical imaging

  • understand the characteristics of images obtained via different mechanisms and modalities

  • roughly understand the applicability of individual modalities to medical diagnostics/therapy

  • demonstrate practical basic image processing skills (filtering, enhancement, restoration, motion, compression, segmentation, registration)

  • discuss applications of image processing techniques to biomedical image processing and choose what types kinds of image processing are suitable for given biomedical applications

  • evaluate whether an image processing system is a good candidate for given biomedical information system

  • keep up with recent advances in image processing and identify possible biomedical applications

 

Course Methodology:

The course will consist of lectures, homework assignments and final projects. Homeworks will include Matlab programming assignments and articles to be read and then discussed in class. Individual students will lead discussions on selected articles. Final projects will be extrapolations of topics discussed within the course (pre-approved by the instructor), with simulations illustrating the results. A final project will include a (journal styled) paper and a 15 minute oral presentation describing the methods and the results.

 

Grading Method:

Main course activities will be evaluated and grading will be as follows:

Class participation: 10%     (according to participation in interactive discussion based on assigned articles)

Homework: 50%        (according to results and written analysis of topic and app. in prog. assignments)

Final project: 40%:               (according to quality of simulation, paper and oral presentation)

top ↑

Data hiding


Course Aims

The advancement in technology offers new solutions, but in the other way, creates new problems as well. Digital media (e.g. audio, images, video, etc.) offers many benefits: it can be stored, duplicated and distributed everywhere in the world, with no loss of fidelity, but in contrary, it can also be manipulated and modified easily, often only with personal computer and appropriate software, and sometimes even unintentionally. While these properties are desirable in general, they can create problems for parties who own digital media and want to distribute it, but at the same time, want to protect it from illegal multiplication and distribution. Digital watermarking has been proposed as a solution for the copyright protection

The goal of the course is to give an introduction to watermarking and hiding messages in audio, image, and video signals, together some techniques for various applications: copyright protection, intellectual property, user identification, and authentication.

 

Course prerequisite:

This course requires basic knowledge of signal processing,

 

Texts:

[1]    I. Cox, M. Miller, and J. Bloom, Digital Watermarking, Morgan Kaufmann Publishers, Inc., San Francisco, 2001.

[2]   S. Katzenbeisser and F. Petitcolas (editors), Information Hiding Techniques for Steganography and Digital Watermarking, Artech House, 2000

 

Course Outline: (Outline syllabus)

Introduction to the problem of data hiding.

Digital watermarking. History of watermarking. Applications and Properties. Models of Watermarking. Methods classification and evaluation. Perceptual Models in digital watermarking. Robustness and security.

Digital watermarking techniques. Content Authentication Steganography and Steganalysis. History of steganography. General concepts and applications of steganography. Multimedia data steganography techniques. Steganalysis.

 

Learning Outcomes

Having successfully completed this unit, you will be able to demonstrate knowledge and understanding of:

·     the basics of digital watermarking

·   the various applications of robust invisible digital watermarking such as copyright protection,  image fingerprinting, image authentication, etc.

·   the technical principles of the various spatial-domain and transform-domain techniques employed for the robust invisible watermarking of images and video

·     the attacks of the watermarking systems as well as methods to countermeasure these attacks.

·     The benchmarks used to evaluate the performance of robust watermarking techniques.

·     the technical description of fragile and semi-fragile watermarking techniques and how they can be used for image authentication

·     the general concepts and applications of steganography

·     the technical principles of steganography techniques

·     state-of-the-art steganalysis techniques that can detect if an image contains an invisible watermark.

Also, upon completion of this course, student should be able to:

·     work in team

·     research, present, and report a selected project within a specified time.

·     think critically, ask questions, and apply problem-solving techniques.

 

Course methodology:

The course concepts will be given by lectures, readings, seminar-like presentations, and projects. Regular lectures will be given in the first half of the semester by the instructor. During this time, students will be reading articles and selecting project topics. In the second half, students will be required to present papers in class and to give progress reports on their projects. There will be several research papers that each student is expected to read, and students will be expected to participate actively in both the lecture and paper discussions. At the end of the semester, each student/team will present their project.

 

Grading Method:

The grading will be based on the student's performance in class participation, paper presentation, and the research project. Specifically:

Class participation: %

Paper presentation: %

Project: %

top ↑

Biometrics and video surveillance


Course Aims

Biometrics is the study of automated methods for the identification or authentication of individuals using biological characteristics (e.g. fingerprint, face, and iris images). Biometrics technology as well as video surveillance methods has the potential to solve key security challenges associated with information systems. This course explores the underlying technology, use, issues, and potential of biometrics and video surveillance for identification and authentication. The selected topics provide an overview of the use of biometrics for identification with an emphasis on its role in information security.

Trough the course the students should develop:
- a knowledge of the most important biometric approaches.
- the capability to select a suitable algorithm / system for a given application context (e.g. physical access control)
- a good understanding of the complex relationships between biometric systems and environmental conditions and their impact on biometric performance.
- the capability to assess the security properties of a biometric system.

 

Course prerequisite:

Basic knowledge of signal processing,

 

Texts:

[1]   John D. Woodward Jr., et al Biometrics, McGrow-Hill Osborne Media 1st edition, 2002
[2]
  A. K. Jain, P. Flynn, A. Ross, " Handbook of Biometrics", Springer, 2007.

[3]   A. Ross, K. Nandakumar and A.K. Jain, Handbook of Multibiometrics, Springer Verlag, 2006.

[4]  A.K. Jain, R. Bolle and S. Pankanti (Eds.), BIOMETRICS: Personal Identification in Networked society, Kluwer Academic Publishers,1999.

 

Course Outline: (Outline syllabus)

Introduction to biometrics. Biometric systems based on fingerprint recognition. Biometric systems based on iris recognition. Face identification and localization in images. Retina recognition methods. Human identification based on gait. Fusion in biometrics (multibiometrics). New trends in biometrics.

Video surveillance – introduction. Video acquisition systems and video quality improvement. Motion detection techniques and object tracking. Recognition humans and their activities in video. Intelligent surveillance techniques.

Barcode and RFID

 

Learning Outcomes

A student who has met the objectives of the course will be able to:

·        Explain concrete biometric models introduced in the course

·        Analyze the suitability of biometric models for a given scenario

·        Use the biometric ingredients of existing systems to obtain a given security goal

·        Indicate the potential limitations biometric ingredients in existing systems for achieving a given security goal

·        Judge the appropriateness of proposals in research papers and text books for a given application

·        Design a biometric solution for a given application scenario

·        The students should be able to achieve the above goals in a group effort while maintaining individual accountability

·        The student should be able to communicate his/her results in a clear and precise manner

 

Course methodology:

The course concepts consist of two parts: the first one when regular lectures will be given by lectures, and in the second one consisting of readings, seminar- presentations, and projects. . In the second half, students will be required to present papers in class and to give progress reports on their projects. There will be several research papers that each student is expected to read, and students will be expected to participate actively in both the lecture and paper discussions. At the end of the semester, each student/team will present their project.

 

Grading Method:

The grading will be based on the student's performance in class participation, paper presentation, and the research project. Specifically:

Class partisipation:30 %

Paper presentation:20 %

Project:50 %

top ↑

Image processing


Course Description

Image processing and analysis represents an exciting and dynamic part of the science of digital signal processing with a significant growth of active applications (remote sensing, technical diagnostics, autonomous vehicle guidance, medical imaging, automatic surveillance etc.), and with a swift progress demonstrated through an increasing number of image processing and analysis software and hardware products on the market.

This course will offer the students the basics of image processing through the coverage of image enhancement, restoration, and segmentation, and of the basics of image analysis through the coverage of shape/texture representation and description, object/texture recognition and motion analysis. The areas of image understanding and 3D-vision will also be briefly introduced.

 

Course Prerequisite:

This course requires basic knowledge of signal processing.

 

Texts:

[1]    Milan Sonka, Vaclav Hlavac, Roger Boyle, Image Processing, Analysis and Machine Vision, Chapman & Hall, 3rd ed., 2008

[2]    Rafael C. Gonzalez, Richard E. Woods, Digital Image Processing, Prentice Hall, 3rd ed., 2007

 

Course Outline: (Outline Syllabus)

·-Digitized image and its properties

·-Data structures for image analysis

·-Image pre-processing: Pixel brightness transformations; Geometric  transformations: Local pre-processing; Restoration

·-Advanced segmentation: Thresholding; Border detection; Advanced optimal border and surface detection approaches

·-Shape representation and description: Region identification; Contour-based shape representation and description; Region-based shape representation and description; Shape classes

·-Object recognition: Knowledge representation; Statistical and syntactic pattern recognition; Recognition as graph matching; Optimization techniques in recognition

·-Matematical morphology: Basic principles and morphological transformations; Topological  processing

·-Texture: Statistical texture description; Syntactic texture descriptions; Hybrid texture description methods; Texture recognition methods applications

·-Image understanding: Image understanding control strategies; Active contour models; Point distribution models; Pattern recognition methods in image understanding; Scene labelling and constraint propagation; Semantic image segmentation and understanding

·-Motion analysis: Differential motion analysis methods; Optical flow; Motion analysis based on detection of interest points; Kalman filters

·-3D-vision: 3D-visioin tasks; Geometry of 3D-vision; Radiometry and 3D-vision

 

Learning Outcomes

Having successfully completed this course, you will be able to:

  • explain the fundamentals of: image formation and acquisition, image representation in the spatial and frequency domains, the respective roles of sampling, quantization, transformation and of the HVS

  • explain the basics of image processing and demonstrate practical skills in: image pre-processing, advanced segmentation techniques, shape and texture representation/description techniques

  • understand shape/texture recognition, and object tracking techniques

  • understand the very basics of 3D-vision and image understanding

  • discuss the applications of image processing, pattern recognition and image analysis to various problem-solving tasks

  • keep up with recent advances in image processing/recognition and identify possible applications

  • evaluate whether an image processing/recognition system is a good candidate for a given information system

  • describe evaluation criteria for image processing/recognition systems

 

Course Methodology:

The course will consist of lectures, homework assignments and final projects. Homework will include Matlab programming assignments and articles to be read and then discussed in class. Individual students will lead discussions on selected articles Final projects will be extrapolations of topics discussed within the course (pre-approved by the instructor), with simulations illustrating the results. A final project will include a (journal styled) paper and a 15 minute oral presentation describing the methods and the results.

 

Grading Method:

Main course activities will be evaluated and grading will be as follows:

Class participation: 10%   (according to participation in interactive discussion based on assigned articles)

Homework: 50%              (according to results and written analysis of topic and approach in programming assignments)

Final project: 40%:           (according to quality of simulation, paper and oral presentation)

top ↑

Image and video compression


Course Description

Due to the widespread application of digital images and video, enhanced quality high-resolution images and/or video are becoming desirable in all cases, and in certain applications they are even essential. Recording, storage, distribution and usage of such images and video are made possible by applying efficient compression techniques.

In this course students will be introduced to basic theory of image and video compression and to a variety of image and video compression techniques and algorithms, as well as compression standards that have widespread use in different fields of application of digital images and video today.

 

Course prerequisite:

This course requires basic knowledge of signal processing.

 

Texts:

[1]   V. Bhaskaran, K. Konstantinides, Image and Video Compression Standards, Kluwer Academic Publishers, 1997.

[2]   Y. Wang, J. Ostermann, Y-Q. Zhang, Video Processing and Communications, Prentice Hall, 2002.

[3]  J.D. Gibson, T. Berger, T. Lookabaugh, D. Lindbergh, R.L. Baker, Digital Compression for Multimedia, Morgan Kaufmann Publishers Inc., San Francisco, CA, 1998.

[4]   J. Watkinson, The MPEG Handbook, 2nd Edition, Elsevier, 2004.

 

Course Outline: (Outline syllabus)

  • Compression fundamentals, Entropy coding

  • Lossless coding: Application of Huffman coding in image compression, JPEG-LS standard

  • Measures of distortion and video quality: Mean squared error (MSE), mean absolute error (MAE), signal-to-noise ratio (SNR), Subjective visual quality, mean opinion score (MOS)

  • Lossy image and video compression: Fundamentals of Rate-Distortion theory, Scalar and vector quantization, Transform coding, Predictive coding, Hybrid coding 

  • Content-based video coding: Shape coding, Texture coding, Object-based coding, Knowledge-based coding, MPEG-4 standard

  • Post-processing techniques for compression artifacts suppression: Compression artifacts detection and estimation, Compression artifacts suppression techniques, H.264 AVC standard

  • Scalable video coding: Quality scalability, Spatial scalability, Temporal scalability, Object-based scalability

 

Learning Outcomes

Having successfully completed this course, the student will be able to demonstrate knowledge and understanding of:

  • basics principles of digital image and video compression

  • contemporary techniques for lossless image compression

  • most frequently used distortion measures and subjective image quality measures

  • the general concept of lossy image and video compression

  • the fundamentals of rate-distortion theory and scalar and vector quantization

  • most frequently used transform coding techniques for image compression

  • the basics of spatial and temporal prediction and prediction techniques used in image and video coding standards

  • hybrid coding

  • the genesis of video coding standards  

  • contemporary video coding standards

  • negative effects of compression and state-of-the-art techniques for compression artifacts suppression

  • basics of context-based video coding

  • the concept of scalability and basic techniques for scalable video coding

 

Also, upon completion of this course, the student should be able to:

  • cooperate with team members of the same or different professional profiles connected to digital image and video storage and distribution, like network and broadcast systems experts; 

  • understand problems, do research and present the results of the research in a given timeframe.

 

Course methodology:

The course material will be covered by lectures, readings, presentations, and projects. Regular lectures will be held in the first two thirds of the semester by the instructor. During this period, while following the lectures, students will be requested to read selected articles and prepare presentations and discussions of the papers read. At the beginning of the last third of the semester the students will choose their project topics. During the last third they will work on their projects and will be requested to give short presentations about project progress. The complete presentation and project report are due to the semester’s end. The students will be expected to actively participate in all activities during the semester. 

 

Grading Method:

The grading will be based on the student's performance in class participation, paper presentation, and research project. Specifically:

Class participation: 20%

Paper presentation: 30%

Project: 50%

top ↑

DSP microcontrollers


Course Aims

Low-end and midrange microcontrollers are nowadays ubiquitous in electronic equipment. They provide versatility and “intelligence” to the equipment performing simple control tasks easing the user’s life. They can also perform the necessary signal processing calculations and/or process control calculations but are limited to relatively slow processes. When high dynamic processes are controlled, such as vector motor control, or high throughput calculations are needed, such as with sensorless or spectral data based control, DSP architectures appear more suitable. DSP microcontrollers are aimed to suit both functions. Since their price is comparable with “classical” midrange microcontrollers, they are expected to penetrate the midrange microcontroller application field.

The objectives of the proposed course are to provide the students with knowledge and hands-on experience in translating DSP and control concepts into real-life software for embedded systems using DSP boards. The proposed course emphasizes designing and implementing software for embedded systems through the programming of a special type of microcontroller, the DSP microcontroller. Laboratory exercises will be based on writing C and assembly language software for selected DSP microcontroller boards that are used in current consumer products, and interfacing them to external devices for test and measurement.

 

Course prerequisite:

This course requires basic knowledge of C programming

 

Texts:

[1]    Magazine articles, DSP microcontroller manufacturer’s manuals

[2]   (TI instead of Freescale) Sen M_Kuo,Woon-Seng S_Gan: "Digital Signal Processors - Architectures, Implementations, and Applications", Prentice Hall, 2004, ISBN: 0-13-035214-4

[3]    Sen M. Kuo, Dennis R. Morgan, “Active Noise Control Systems: Algorithms and DSP Implementations”, Willey, ISBN: 978-0-471-13424-4

[4]     Other Control Topics…

 

Course Outline: (Outline syllabus)

Introduction to DSP microcontrollers, general characteristics, examples of DSP microcontroller typical applications and families.

Detailed architecture of a selected DSP microcontroller, instruction set, development environment.

Overview of peripherals.

New advances in configurable microcontrollers.

Topics in control, optimal digital control, high performance digital control, active noise control (…).

Application examples, libraries; control (…) and DSP capabilities.

 

Learning Outcomes

Having completed this unit, students will be able to demonstrate:

·   Knowledge of a DSP microcontroller architecture and it’s influence on control and digital signal processing functions

·   Experience using equipment and tools commonly used in industry and/or experimental laboratory

·   Understanding of the process of converting ideas and algorithms into real-life working hardware with real-life signals

 

Also, upon completion of this course, students should be able to:

·   work in team and individually

·   prepare for, perform and report experimental work

·   self-organize to acquire knowledge, implement it and track the progress through a real-life project(s)

 

Course methodology:

The course will be given by lectures, readings, discussions, laboratory experiments and mini-project(s). After the general introductory lectures given by the instructor, students will be studying chapters from the text-book(s), datasheet and manuals and preparing written work plans for performing the assigned experiments. At the same time they will be exposed to certain examples and problems where control and sensor signal processing is needed to choose from for their project assignment. The work plans will consist of homework assignment as theoretical preparation and practical work preparation, and will be discussed and graded in class prior lab. experiment. After the lab. each student will prepare a report. After achieving the basic level of mastering the required tools and techniques, students will start working on their projects and will be giving regular reports on their progress that will be discussed and graded in class. The course will end with a final report, presentation and demonstration of their work.

 

Grading Method:

The grading will be based on student’s performance in written work plans and discussions, laboratory work, reports preparation and discussion and project report presentation and demonstration. Specifically:

Work plans: 20%

Reports: 20%

Laboratory: 20%

Project: 40%

top ↑

Digital audio and speech analysis and processing


Course Description

Progress in digital audio and speech processing has been phenomenally rapid. It is employed in recording and storing music and speech signals, for sound mixing and production of digital programs, in digital audio broadcasting and digital television, in cellular telephony as well as in consumer products like CDs, DVDs and PCs.

This course provides the basis of an advanced course in digital audio signal processing and is directed at students studying engineering, computer science, physics but also for professionals looking solutions to problems in audio signal processing in studio engineering, consumer electronics and multimedia.

 

Course prerequisite:

Prior basic knowledge of acoustics, systems theory and digital signal processing are taken as a prerequisite.

 

Texts:

Basic:

[1]    Udo Zolzer, Digital Audio Signal Processing, John Wiley & Sons Ltd. , 1997.

[2]    R. Rabiner and R.W. Schafer, Digital Processing Of Speech Signals, Prentice Hall, 1978

[3]    Sadaoki Furui, digital speech processing, synthesis and recognition, Marcel Dekker Inc., 2001

 

Additional:

[1]   Mark Kahrs, Karlheinz Brandenburg ed., Applications Of Digital Signal Processing To Audio And Acoustics, Kluwer Academic Publishers, 2002

[2]     Udo Zolzer, DAFX - Digital Audio Effects, John Wiley & Sons Ltd., 2002

[3]    Saeed V. Vaseghi, Advanced Digital Signal Processing And Noise Reduction, Second Edition. (chs.1, 2, 5, 8.5, 11.4, 12.6, 14), John Wiley & Sons Ltd , 2000

[4]     Chou, Juang, Pattern Recognition In Speech And Language Processing, (chs.1, 4, 5, 7, 8), CRC Press, 2003

 

Course Outline: (Outline syllabus)

Introduction to digital audio transmission system and storage, sampling rate audio conversion, sigma/delta modulation, dither, noise shaping. Data compression fundamentals. Principal characteristics of audio and speech. Compression techniques. Psychoacoustic models. Speech analysis and coding techniques.

Audio and speech synthesis: Waveform synthesis, TTS. Principles of speech (speaker) recognition. Digital audio effects and synthetic music.

 

Learning Outcomes

Having successfully completed this unit, you will be able to demonstrate knowledge and understanding of:

·   the basics of digital audio signal processing hardware and algorithms for processing digital audio and speech signals.

·   the various coding algorithms applied in digital audio recording and storage.

·  the technical principles of the various techniques employed in analysis and synthesis of audio and speech, digital audio effects and synthetic music.

·   the general concepts and applications of speech-to-speech synthesis.

·   the technical principles of speech and speaker recognition techniques.

Also, upon completion of this course, student should be able to:

·   work in team

·   explore technical papers, write reports and present and a selected project within a specified time.

·   think critically, ask questions, and apply problem-solving techniques.

 

Course methodology:

The course concepts will be given by lectures, readings, seminar-like presentations and individual (or team)  projects. Regular lectures will be given in the first half of the semester by the instructor. During this time, students will be reading articles and selecting project topics. In the second half, students will be required to present papers in class and to give progress reports on their own projects. There will be several research papers that each student is expected to read, and students will be expected to participate actively in both the lecture and paper discussions. At the end of the semester, each student/team will present their project report.

 

Grading Method:

The grading will be based on the student's performance in class participation, specific paper presentation, and the research project report. Specifically:

Class participation: 10%

Paper presentation: 30%

Project: 60%

top ↑

Application of neural networks in signal processing


 

top ↑

Multimedia technologies


Course Description

Multimedia has become a widespread and indispensable part of modern computer technology.

The goal of this course is to introduce students to the principles and technologies of multimedia systems, as well as to provide them with hands-on experience in this area. Issues on effective ways to represent, process, communicate and retrieve multimedia data, such as computer graphics, digital audio, image and video, will be addressed.

 

Course prerequisite:

This course requires basic knowledge of signal processing.

 

Texts:

[1]   Ze-Nian Li, Mark S. Drew, FUNDAMENTALS OF MULTIMEDIA, Prentice Hall, 2003.

[2]   Y. Wang, J. Ostermann, Y-Q. Zhang, VIDEO PROCESSING AND COMMUNICATIONS, Prentice Hall, 2002.

[3]   Ralf Steinmetz, Klara Nahrstedt, MULTIMEDIA SYSTEMS, Springer-Verlag, 2004.

 

Course Outline: (Outline syllabus)

  • Introduction to multimedia: What is multimedia, Historical overview

  • Fundamentals of computer graphics

  • Fundamentals of digital audio processing: Audio compression

  • Fundamentals of digital image and video processing: Image analysis and segmentation, Image and video coding

  • Multimedia communications: Computer and multimedia networks, Error control, Multimedia streaming, Quality of service, Media-on-Demand

  • Content-based retrieval in digital libraries: Retrieving multimedia data (sound, music, graphics, image and video), Principles of image and video databases, MPEG-7 standard

  • Advanced multimedia techniques

  • Sound, speech and gesture recognition

 

Learning Outcomes

Having successfully completed this course, the student will be able to demonstrate knowledge and understanding of:

  • the basics of computer graphics, digital audio processing and digital image and video processing;

  • technical principles of multimedia communications;

  • information transfer in computer and multimedia networks;

  • methods and techniques for error control in multimedia systems;

  • establishing quality of service in multimedia networks;

  • interactivity in multimedia systems and media-on-demand;

  • the general concept of content-based retrieval;

  • techniques and algorithms for multimedia data retrieval;

  • basic principles of video databases and supporting standards;

  • state-of-the-art multimedia techniques mainly for sound, speech and gesture recognition.

 

Also, upon completion of this course, student should be able to:

  • work in a team consisting of experts of different professional profiles like DSP experts, network and database experts; 

  • understand the problems concerning the interoperability of different systems and the effects of their interconnection;

·     conduct research in an interdisciplinary environment and present the results of the research in a given timeframe.

 

Course methodology:

In this course the students will attend lectures during first two thirds of the semester. In the same period they will be assigned selected articles for reading, for which they will prepare presentations and discussions. All students will be required to attend presentations and participate in discussions. At the beginning of the last third of the semester the students will choose their project topics. During the last third they will work on the project and will be requested to give short presentations about project progress. Team work will be supported, and the project can be assigned to teams of students. The complete presentation and project report are due to semester’s end. The students will be expected to actively participate in all activities during the semester. 

 

Grading Method:

The grading will be based on the student's performance in class participation, paper presentation, and research project. Specifically:

Class participation: 20%

Paper presentation: 30%

Project: 50%

top ↑

  
Copyright © 2008 DSP Group FEIT, Skopje, Macedonia