Department of Mathematical Sciences Mathematical of Department Faculty of Faculty Computing Sciences

Full-time European language

Computing Sciences and Engineering and Sciences Computing Part-time Study abroad Sandwich Joint/Combined Industrial placement Distance learning Work experience Colleges

Digital Signal and Image Processing MSc

Location: City Campus, Leicester Who is the masters degree for? Duration: 1 year full-time; 2–4 years part-time The programme is an advanced modular programme This programme is organised by the Department of for science and engineering graduates with a numer- Mathematical Sciences and the Centre for Digital ate academic background. It is concerned with Signal and Image Processing at De Montfort Univer- training students in the art of professional software sity, Leicester. engineering for designing, building and executing individual signal and image processing modules Digital Signal Processing (DSP) and Digital Image (forming part of a developing software library) or Processing (DIP) are at the forefront of information integrating signal processing systems designed technology. They are the basis for a broad range of specifically for a well defined problem. applications including medical imaging, remote sensing, digital communications, geophysical pros- No former software engineering background is pecting, space exploration and robotics to name but a required, but the programme does assume a basic few. DSP and DIP are interdisciplinary subjects which literacy in graduate mathematics and computing. utilizes a wide range of mathematical and computa- Typically, graduates of this programme will find tional methods. There is currently a major shortage of employment as programmers and software engineers, specialists in the design, implementation and integra- technical advisers and systems analysts in the tion of software for DSPs and their component sub- software industry - in both the private and public systems. This programme is designed to supply the sectors. Alternatively, their new skills may facilitate necessary training for graduates to enter industry with their development in existing professional careers. programming and software engineering expertise for digital signal and image processing. Programme

As microprocessor CPUs become increasingly more powerful, there is less demand for designers of specialist DSP hardware units and greater need for software engineers to design modules and objects for execution on low- and high-level platforms. This programme provides a comprehensive coverage of those areas of mathematics, engineering and comput- ing which are necessary for engineers to model, process and analyse digital signals and images.

Preparation for Life By emphasising the practical application of the Signal Analysis theoretical concepts and the central role of software Covers the mathematical methods used for modelling engineering, the programme aims to equip students digital signals and images and designing algorithms with knowledge and skills that will make them rapidly to process them. The syllabus includes a discussion capable of useful employment without extensive of the classical and generalised Fourier transforms, further training. Whilst all necessary theory is cov- the Hilbert transform and quadrature detection, ered, the emphasis throughout is on understanding convolution and correlation, the sampling theorem, and developing the ability to apply and adapt theory to statistical analysis of signals, optimisation methods, real engineering problems. The modules have been wavelets and fractal signals. designed to train students to design complex DSP Digital Signal Processing and image processing systems and to provide Introduces students to the operation, design and knowledge on how these systems are constructed. applications of various DSP techniques and to the Extensive use is made of MATLAB and in particular, design methodology’s for researching and developing the DSP and DIP Tool-boxes. DSP systems. This is based on the design of hard- ware and software modules and the use hierarchy. Programme modules Object oriented design methods are also considered. The modules delivered are summarised below. Elements of the syllabus include the design of recursive and non-recursive digital filters, frequency Programming and Software Engineering domain data processing, time-frequency analysis, This module provides a ‘hands-on’ programming digital simulation, extraction of signals from noise, approach to software engineering. It covers struc- Bayesian estimation methods and non-linear filter tured and modular programming in C, object oriented design. Instruction is given on the design, program- design using C++, the definition and specification of ming and testing of various DSP algorithms in C/C++ a software system, development phases, testing to help students construct their own software library techniques, reliability and maintenance. making use of the MATLAB DSP toolbox as a programming workbench and prototyping environ- Computational Methods ment. In designing algorithms to process digital signals and images, it is necessary to make provision for the Digital Image Processing numerical solution to various types of linear systems This module extends the material covered in the DSP of equations. These systems arise from the analysis module and explores the algorithms available for of digital signals and in particular, the design of digital image restoration and reconstruction, methods of filters. This module discusses the different classes of image enhancement, segmentation techniques, solution that are available and addresses the linear transformation methods, noise reduction, image eigenvalue problem and singular value decomposi- compression and the statistical analysis of images. In tion. each case, real-life problems are explored covering a range of applications and instruction is given on the mathematical models that are used and the circum- stances under which these models and the algorithms derived from them can be employed. As in the DSP module, students develop a library of software in C/ C++ for image processing using the MATLAB toolbox as a prototyping environment to investigate different techniques for DIP. Image Understanding How is the Programme organised Covers the underlying physical principles and The programme is composed of three Semesters each mathematical models which are common to imaging of which is 4 months in duration. The taught part of systems. It is shown how an understanding of the the programme is confined to the first two semesters, ‘physics’ of an imaging system can provide a suitable starting in September and students are initially mathematical model for an image which leads to assessed by a combination of examinations, course- methods of processing and analysing an image. The work and/or minor project work. A crash course in module discusses techniques for simulating images technical English is available prior to the start of the derived from passive and active imaging systems, the programme in August and September together with principles and application of Fourier optics and the short courses in basic computing as required. Each Abbe theory of imaging. It also includes a study of the module consists of a 30 hour lecture course with applications of electromagnetic and acoustic imaging computing laboratory sessions. Programme notes and techniques. other supplementary material is issued for each Intelligent Systems module together with appropriate software. Examina- Considers the specification and construction of tions are held at the end of each semester. In addition techniques of Artificial Intelligence (AI) for signal and to formal lectures, students are encouraged to attend image analysis. Emphasis is placed on the application a series of seminars run by the Department of Math- of Fuzzy Logic and Artificial Neural Networks (ANNs) ematical Sciences and the Centre for Digital Signal for adaptive control and computer vision respectively. and Imaging Processing at De Montfort University, in The syllabus includes adaptive fuzzy inference which personnel from both academic and industrial systems, the application of ANNs for pattern recogni- organisations present research papers and provide tion, AI in process engineering and knowledge-based information on software products and their applica- systems. tions to signal and image processing. A number of short courses are held throughout the year on emerg- Graphical User Interfaces ing areas of importance in DSP which are not covered Introduces the methods of design required in the in the masters programme but are open to all stu- development of Graphical User Interfaces (GUIs) for dents. A block diagram illustrating the basic organisa- ‘driving’ complex signal and image processing tion of the programme is given in Table 1 on the back systems and libraries. Both X-windows and NT- of this brochure. windows systems are used and students are intro- duced to a range of state-of-the-art Computer Aided Financial support Software Engineering Tools for developing GUIs on A number of studentships provided by a variety of both high- and low-level platforms. The syllabus manufacturing industries and software houses are covers the following areas: X-windows, Motif pro- available. In addition, a limited number of scholarships gramming, top-down design, graphics standards, X- are available which are awarded on a competitive designer, visual C++ development systems, Java and basis. J++ development tool-kits.

Flexible part-time study Research Project The project can be taken either at De Montfort Flexible arrangements are possible for students University or on location in industry, typically in a wishing to take the programme on a part-time basis. software company. The research project provides the The usual scheme is to take the programme over two opportunity to develop, to demonstrate and to years, attending four modules each year. The assigned appraise the knowledge and skills acquired from the practical work for each module and the associated programme in the solution of real practical problems deadlines can be arranged to suit students other subject to typical commercial constraints. All research commitments. Those proceeding to project work are projects are concerned with the pursuit of practical allowed 12 months from the last written examination knowledge and its application to real-life problems. before handing in their dissertations. Case Study: Table 1: Nominal class hours Ursula Augsdorfer 1st Semester Ursula Augsdorfer • Programming and Software Engineering 30 studied Mathematics at • Computational Methods 30 the Berlin Technical • Signal Analysis 30 University from 1990 to • Digital Signal Processing 30 1996 specialising in Total lecture hours: 120 the area of mathemat- 2nd Semester ics for information • Digital Image Processing 30 technology. She studied for her masters degree in • Image Understanding 30 Digital Signal and Image Processing in order to • Graphical User Interfaces 30 further her education in the area of information and • Intelligent Systems 30 communications technology. Her research project Total lecture hours: 120 was on the simulation of medical ultrasound images. ‘The MSc provided me with a broad range of the theoretical and practical aspects of signal and image 3rd Semester processing with particular emphasis on the necessary Research project working with a research group at programming skills and software engineering tech- De Montfort University Leicester or in industry. The niques. The programme has significantly helped me to research groups available include the Centre for develop my career prospects for an academic and/or Digital Signal and Image Processing based in the industrial research career.’ Ursula is currently Institute of Simulation Sciences and the Centre for studying for a PhD working in the Institute of Simula- Modern Optics based in the Science and Engineering tion Sciences at De Montfort University. She is Research Centre. Descriptions of available projects researching a technique for simulating the non-linear with expected deliverables are issued to students dynamical behaviour of an imploding shock wave early in the second semester. which is believed to be associated with the phenom- enon of sonoluminescence. The simulation involves developing numerical solutions to the Reyleigh- Plesset equations and a fully 3D visualisation system for analysing the behaviour of this non-linear effect in Further information real time. Application forms and further details are available from the Postgraduate Admissions Tutor:

The programme has been designed to be suitable for Tel: Leicester +44 116 257 7699 the in-service training of existing scientists and Fax: Leicester +44 116 254 1891 engineers as well as those seeking a change of Email: [email protected] career. Such applicants are invited to discuss their www: http://www.cms.dmu.ac.uk/ particular circumstances with the programme director Courses/course-list.html Professor J M Blackledge tel: +44 (0) 116 257 7473, FOR GENERAL ENQUIRIES e-mail: [email protected]. 0645.45 46 47 LO-CALL

SEE OUR WEB SITE AT http://www.dmu.ac.uk

H1270

© The Marketing Division, De Montfort University.June 1999. Design CETD print The Printing Centre De Montfort University. Right of revision: this information is issued without prejudice to the right of the University authorities to make such modifications to the matter dealt with as the University authorities consider necessary without prior notice.