Computing Science

University of Stirling

Autumn 2002 Seminars

The Department of Computing Science and Mathematics presents the following seminars. Unless otherwise stated, seminars will take place in Room 4B94 of the Cottrell Building, University of Stirling from 15.00 to 16.00 on Friday afternoons during semester time.

If you would like to give a seminar to the department in future or if you need more information, please contact the seminar organisers, either Julie Hodgkin (Phone 01786-467-446, Email or David Cairns (Phone 01786-467-445, Email

5th September 2002
 Plamen L. Simeonov,  Technology University of Ilmenau
Viator, the Wandering Network approach, defines a new type of communications architecture chracterized by:

• flexible, multi-modal specialization of network nodes as virtual subnetworks
• mobility and virtualization of the net functions in hardware and software
• self-organizing topology-on-demand.
Network elements can contain several exchangeable modules capable of executing diverse network functions in parallel. They can be invoked,  transported to or generated in the nodes upon delivery of mobile code about the node’s behaviour.
20th September 2002
Size is important: Data compression for mobile applications
John Wilson, Department of Computing and Information Sciences, University of Strathclyde
Conventional database systems rely on disk-based technology. This provides large-scale storage at the expense of limited performance by comparison with that available in RAM-based systems. Main-memory resident databases are however limited by the relatively meagre provision of RAM in most systems. Dictionary based data compression can be used to squeeze data and partially overcome the problems of limited capacity. It also acts as a basis for fast query processing and efficient data distribution. This talk describes a technique that provides a compact data representation to enable queries to be carried out in the compressed domain. Further we consider the kind of application that could benefit from this technology in the light of the emerging potential for mobile computing.

27th September 2002
Distributed & parallel processing of motion parallax depth cues by dynamic mapping
Markus Dahlem, Computational Neuroscience, Department of Psychology, University of Stirling
One of the main problems in vision is the three-dimensional reconstruction of a static scene from a sequence of images taken by an observer. The observer can move around and/or changes direction of gaze. This will lead to complex changes in brightness, called optical flow, on the projection plane. As I will show, dynamic retino-cortical mapping can be used to reduce the originally two-dimensional optical flow field on the retina to an one-dimensional flow in a "higher" area of visual cortex and thus simplifying depth cues in the images.  My goal is to build an artificial visual system in close approximation of existing brain functionality, i.e. cortical sensor maps, and distributed & parallel neuronal operations.

4th October 2002
GIS: what's it all about and what can we do with it?
Sandy Winterbottom, Department of Environmental Science, University of Stirling
The aim of this talk is to provide an introduction to GIS and it's potential applications.  The talk will be illustrated with examples from Environmental Science but also introduce other potential application areas.  Software and data availability within the University will be discussed.

11th October 2002
The Rules of Sailing Races for Hand-Held Devices
Ken Turner, Department of Computing Science and Mathematics, University of Stirling

The motivation is given for having computer support of the rules for sailing races. SailRule is a freely available program intended to analyse and improve performance in applying the racing rules. A brief overview is given of sailing terminology and racing rules. It is argued that a useful program for the racing rules should run on hand-held devices. The program should support an archive of rule scenarios, race training, self-learning of the rules, and analysis of rule disputes. The SailRule program has been implemented using the SuperWaba programming environment for hand-held devices. The user interface is described for the rules program. An explanation is given of the principles behind formalising and codifying the racing rules so that they can be efficiently implemented. Examples are given of how the program represents and analyses rule scenarios.
18th October 2002
Selecting Test Cases from Formal Proof
Jeremy Bryans and Savi Maharaj, Department of Computing Science and Mathematics, University of Stirling
Given a formal proof of the correctness of an abstract model of some program, we wish to generate test cases which can be used to verify a concrete implementation of that model.   Our hypothesis is that a rigorous method of selecting test case data from the abstract model will be especially effective at locating  flaws within concrete implementations of the model.  We will discuss the context of this research, and report on our current work.
25th October 2002
The Stable Marriage Problem and Constraint Programming (SM & CP)
Patrick Prosser, Department of Computing Science, University of Glasgow
In the stable marriage problem we have n men and n women. Each man ranks each of the women into a preference list. The women do the same with the men. The problem is then to marry men to women such that marriages are stable, i.e. there is no incentive for divorce. In 1962 Gale and Shapley proposed a polytime algorithm for this problem (actually, linear in the size of the problem). Recently we (myself, along with Ian Gent, Barbara Smith, David Manlove, and Rob Irving) have shown that with a constraint programming representation of the problem, arc-consistency processing achieves the same results as the Gale Shapley algorithm. What's more, one of our constraint programming representations results in optimal complexity, the same as Gale Shapley.

In this talk I will introduce the stable marriage problem, present Gale & Shapley's algorithm, and show how we can represent this as a constraint program. I will also present the results of an empirical study of stable marriage problems with ties and incomplete lists, a problem recently shown to be NP-Complete.
1st November 2002 - Mid-Semester Break.  No Seminar today

8th November 2002
Biosignal Analysis: How Much Should You Expect from It?
Minija Tamosiunaite, Department of Applied Informatics, Vytautas Magnus University, Lithuania

Computer scientists sometimes spend their lives working on "universal" signal analysis techniques. However, each problem in biosignal analysis has its own specificity, which makes those theoretically sound methods only of limited use. The viewpoint will be introduced through examples from electrocardiogram (ECG) analysis. Linear modeling, neural networks, chaos theory-based methods will be concerned. New trends in ECG analysis will be revealed.
15th November 2002
ICU-Talk / Interactive Visualisation of a Volume Rendered Virtual Colonoscopy Using a Desktop PC
Ian Ricketts, Department of Applied Computing, University of Dundee
            This talk will be divided into 2 parts.
Developing communication software for people with special needs can be a challenging task. Special attention needs to be given to interface design, the cognitive load placed on the user and to make systems intuitive and transparent. This approach to software development was used to develop a specific communication aid for intubated patients in the intensive care unit (ICU). These patients are temporarily unable to communicate. The literature states that current low tech communication solutions such as alphabet charts or picture boards are often unsatisfactory, slow and frustrating for the patient and nursing staff. ICU-Talk is a prototype computer based communication aid which has been developed to meet the specific needs of the ICU patient and ICU environment.
This presentation will discuss the challenges faced when developing and testing the ICU-Talk prototype, the results from this project and future developments.
Interactive Visualisation of a Volume Rendered Virtual Colonoscopy Using a Desktop PC
Colorectal cancer is the second most common cause of cancer deaths in industrialised societies. Early detection relies on population screening and significantly increases the chances of survival. Virtual colonoscopy (VC) is a minimally invasive screening technique for colorectal cancer. We have developed a fast three-stage method for volume rendering VC. In the first stage the lumen boundary is identified and using a thickening technique a layer from lumen to the colon wall is then extracted. In the second stage we further reduce the time to render each frame by use of a visibility determination technique. The third stage is an object order volume rendering using the potentially visible voxels belonging to the extracted layer. Shear-warp factorisation provides efficient access to these voxels.
Results were obtained using a synthetic colon on a desktop personal computer (PC). Pre-processing to extract the colon wall and to determine the visibility took 50 secs. The first stage of the method removed approximately 95% of the data and the visibility determination stage eliminated a further 70%. Visualisation of each frame of the VC required approximately 1.5% of the original data set. This was rendered at 5 frames per second, which would enable a clinician to interactively navigate the virtual colon on a desktop PC.
22nd November 2002
Incompatibilities between communications services in a deregulated market
Mario Kolberg, Department of Computing Science and Mathematics
Triggered by the deregulation of the communications market, it is expected that the number of communication services and indeed the number of service providers will increase dramatically in the near future. Moreover these services will interwork resulting in the service interaction problem. This is a problem where the goals of the individual services clash. It is essential to cope with the problem in a multi-vendor environment as it may substantially delay service deployment and form a serious obstacle to rapid service provisioning.
The talk will describe the expected problems and explain a novel solution to address them. The approach can be employed as an off-line technique where it acts as a filter. However the approach has been adapted to operate at run-time both within the intelligent network and within the SIP VoIP architecture. SIP is a session control protocol for Voice over IP sessions.

However communication services are only one aspect of networked services. The second part of the talk will concentrate on service interactions between networked "intelligent" appliances within HomeNetworks. Networked appliances attract an increasing interest from industrial players. Common, widely referenced examples of networked appliances include video cameras, the internet alarm-clock, fridges, and entertainment devices such as TVs and VCRs. The appliances are controlled by software and will interwork. Hence as in traditional communications, the service interaction problem exists. The talk will present a taxonomy of the problem in the appliances domain and an outline solution will be discussed.

The research covered in this talk reflects my PhD thesis.

29th November 2002
Come and meet Toffoli Gates
Stephan Reiff-Marganiec, Department of Computing Science and Mathematics
Toffoli Gates is not Bill's younger sister. In fact she is not related to the MicroSoft dynasty at all, but there is a certain relationship to Fredkin. It turns out that both (Toffoli and Fredkin) are Logic gates with a very specific porperty: they have as many outputs as they have inputs. They seem to form a suitable basis for qualtum computing. I will present some of the basic ideas and discuss why they are seen to be useful.

This is not related to my research. I came accross these gates in a computer magazine, and simply found them interesting.
4th  December 2002  - Room 2B74
Building Networked Appliances for my Dad
Dave Marples,  Honorary Professor, Department of Computing Science and Mathematics
Computational devices are becoming ever more pervasive in our lives - already our cars, televisions, video recorders and even our toasters contain microcontrollers to enhance their functionality. As we can achieve more and more at cost points that are compatible with consumer electronics the temptation is to push more and more functionality into these everyday devices, but in this presentation it is argued this we must resist this temptation, otherwise the devices will become too complex for the average user to fully exploit, just as happened with the general purpose PC today.

Previous Seminar Series

Spring 2002 Seminars
Autumn 2001 Seminars
Spring 2001 Seminars
Autumn 2000 Seminars

To Computing Science and Mathematics Seminars (Current Series)
To Computing Science and Mathematics Research
To Computing Science and Mathematics Home Page