UCL DEPARTMENT OF COMPUTER SCIENCE Information for Students
	Home	Admissions	Students	Careers	Research	Business	People	Help

Text size A A A A A

| STUDENTS > Medical Scientific Computing |

Medical Scientific Computing

Note: Whilst every effort is made to keep the syllabus and assessment records correct for this course, the precise details must be checked with the lecturer(s).

Code:	3053 (Also taught as: GC08)
Year:	3
Prerequisites:	Successful completion of years 1 and 2 of the Computer Science programme
Term:	2
Taught By:	Andrew Todd-Pokropek (100%)
Aims:	The aims of the course are to provide the student with enough information that they may be able to understand the uses of computers for processing in particular pictorial information in the medical context. The bulk of the material is concerned with image processing, with a medical flavour. Thus some basic material is considered such that the student can understand how to acquire data from a variety of system as used in medicine, for example how such data should be sampled. The tools for data manipulation are considered in a variety of applications, for example Computerised Tomography, Nuclear Magnetic resonance, etc, which also involve understanding a number of associated processes such as tomographic reconstruction. These tools are placed in the context of real medical applications such that the student should be able to understand not only what can be done, but how to assess the value of what has been attempted. Finally, the student should be able to understand how such specific applications of computing in medicine (/medical imaging) may be linked together and connected to more general information handling systems, how data may be transmitted, archived, and managed.
Learning Outcomes:	The student should be able to understand and code a number of simple image processing algorithm and therefore have a background suitable such that they would be eventually suitable for employment in this application domain.

Content:

Introduction	What sort of data are we handling, and why? Place of computing within medicine:- Analogue v. digital techniques, Diagnosis v. therapy. Overview:- acquisition, processing, display, interpretation, Signal v. Image Processing
Basic Tools	Sampling, and the sampling theorem, (in 1-D, 2-D ...) Aliasing Hence- the Fourier transform and the convolution theorem Linear and stationary system Simple filtering:- Smoothing, inverse filter, unsharp masking, Butterworth, the matched filter and the Wiener filter. Matrix representation Segmentation:- Differentiation and Edge detection Elements of pattern recognition:- Skeletonisation, clustering, line and shape detectors, Elements of mathematical morphology
Further tools	Basic numerical analysis, and error propagation Non-linear systems and filters Regions of Interest and flood filling. Functional images Data structures: quad and oct-trees Multi-resolution methods Advances in segmentation and feature detection Image registration and fusion, head hat, and voxel based methods Evaluation of algorithms and procedures
Display	Graphics v. Image display Interpolation and Lookup tables Histogram equalization and Variable histogram equalization Contours and 2-D graphics 3-D display Architecture for display workstations
Tomography	Longitudinal v. transaxial Statement of Radon and Inverse Radon transforms Forward and Back projection Interative methods: ART, MLEM Filtered backprojection Window functions, noise and artefacts Optimization of reconstruction algorithms Architecture of computer systems for tomographic reconstruction
Applications	Nuclear Magnetic Resonance and basic MRI image formation Digital Radiography and photostimulable plates X-ray Computerized Tomography and spiral CT Radiotherapy treatment planning Ultrasonography (A, B, M and Real time Nuclear Medicine, and SPECT ECG, EEG, EMG signal processing
Image networks and PACS	Data compression:- Runlength coding, transform coding, JPEG DPCM, Huffman and Lempel-Ziv. Architecture and Networks: data-rates etc DICOM and other standards Object models for PACS and reporting Interfaces to PAS and hospital management systems
Interpretation	ROC curve analysis Resolution requirements How to integrate digital techniques into an analogue environment Social (and economic) implications
Some clinical problems, and examples	Cardiac function Blood flow by NM, NMR, DR, US Masked subtraction and stenosis detection Brain analysis (fMRI, DWI) Mammograms and Computer Aided Detection/Diagnosis Lung and liver nodule detection and classifiication Intervention and computer assisted surgery Intensive care and healthcare in the home

Method of Instruction:

Lecture presentations with associated coursework problems.

Assessment:

The course has the following assessment components:

Written Examination (2.5 hours, 75%)
Coursework Section (3 pieces, 25%)

To pass this course, students must:

Obtain an overall pass mark of 40% for all sections combined

The examination rubric is:
Answer 3 questions out of 6.

Resources:

Image Processing, Analysis and Machine Vision (2nd edition), M. Sonka, V. Hlavac, R Boyle. Chapman and Hall, 1999, ISBN 0-534-95393-X

Handbook of Medical Imaging Volume 1,Physics and Pyschophysics, J. Beutel, H. Kundel, R Van Metter. SPIE Bellingham, 2000, ISBN 0-8194-3621-6

Handbook of Medical Imaging Volume 2, Medical Image Processing and Analysis, J. Fitzpatrick, M. Sonka. SPIE Bellingham, 2000, ISBN 0-8194-3622-4

Additional information

This page last modified: 26 May, 2010 by Nicola Alexander