IT Technical Report Series for the Year 2001

[ 2009 ( 1 ) ][ 2005 ( 3 ) ][ 2004 ( 12 ) ][ 2003 ( 9 ) ][ 2002 ( 16 ) ][ 2001 ( 12 ) ][ 2000 ( 11 ) ][ 1999 ( 15 ) ]

Technical Reports

Deconvolution of drizzled and rotated HST/WFPC2 images: Faint-object photometry in crowded fields

Technical Report No. : NUIG-IT-190201
By : R. Butler, A. Golden and A. Shearer

We have further adapted our deconvolution-based HST/WFPC2 reduction techniques, which our simulations have shown to be very effective in crowded fields (Butler 2000), to take advantage of the extra resolution afforded by ``dithered'' observations. This appears to be the first attempt at non-linear restorations of data which had already been sub-sampled and combined in a linear manner using the STSDAS/drizzle software, thus attaining a net subsampling factor of 4x4. The motivation was a search for the optical counterpart of the millisecond pulsar PSR B1821-24 in the globular cluster M28; the technique is illustrated by our photometric search for candidates in the radio-derived error circle, in archival HST/WFPC2 images in both the F555W & F814W bands of the core field of M28.

Download the Full Paper

HST Photometry of Dense Globular Cluster Fields: The Advantages of Deconvolution

Technical Report No. : NUIG-IT-160201
By : R. Butler

It is shown that in very crowded fields, such as the cores of globular clusters, aperture photometry of subsampled-deconvolved HST/WFPC2 images gives statistically better results than any of the following DAOPHOT-II reductions of the original data: aperture photometry, profile-fitting photometry, and the hybrid method of aperture photometry on neighbour-subtracted images. Furthermore, the deconvolved images also provide the basis for improved star detection, using otherwise conventional means. These conclusions are developed and illustrated using both real HST/WFPC2 data of the Milky Way bulge globular cluster NGC 6293 in the F555W and F814W bands, and realistic simulations of a globular cluster observed under similar circumstances. The results are a consequence of exploiting the nature of the HST/WFPC2 point-spread function (PSF): undersampled, spatially varying and with high-frequency structure in the wings, but determinable to high accuracy by subsampled spatially-varying modelling of first-guess Tiny Tim grids followed by empirical improvements

Download the Full Paper

Parallel image restoration with spatially variant point-spread-function - Description and first clinical results

Technical Report No. : NUIG-IT-140201
By : A. Shearer, G. Gorman, T. O'Doherty, W. van der Putten, P. McCarthy and L. Jelen

In this paper we present a parallel code which performs an iterative image deconvolution using either a spatially invariant point spread function (SI-PSF) or a spatially variant point spread function (SV-PSF). The basic algorithm (Richardson-Lucy maximum likelihood iterative procedure) is described as well as a description of the parallel implementation. Applications and results in the area of medical x-ray imaging are discussed.

Download the Full Paper

Development of Extranet Technology for Investment Banking Clients

Technical Report No. : NUIG-IT-200101
By : T. Hall and D. Chambers

Drucker[1] predicts that if businesses fail to deploy their products and services effectively and expeditiously over the Web, they will, ultimately, "be killed by e-commerce". The Internet is revolutionizing business, to the extent that the Internet is generating new imperatives for the way in which business is practically conducted[1]. In this paper, we identify and describe how the Internet is redefining the investment bank-client relationship, and investigate the effectiveness of the extranet and related technologies in enabling investment banks to fulfil their new, Web-borne business-imperatives, and facilitate effective online interaction with their clients.

Download the Full Paper

Combining Feature Selection and Neural Networks for Solving Classification Problems

Technical Report No. : NUIG-IT-130601
By : P. O'Dea, J. Griffith and C. O'Riordan

This paper presents an approach to solving classification problems by combining feature selection and neural networks. The main idea is to use techniques from the field of information theory to select a set of important attributes that can be used to classify tuples. A neural network is trained using these attributes; the neural network is then used to classify tuples. In this paper, we discuss data mining, review common approaches and outline our algorithm. We also present preliminary results obtained against a well-known data collection.

Download the Full Paper

Genetic Algorithms, their Operators and the NK Model

Technical Report No. : NUIG-IT-150601
By : S. Hill and C. O'Riordan

This paper outlines the operators and workings of Genetic Algorithms, and Kauffman's NK model. To analyse the performance of genetic algorithms and their operators the fitness landscape is crucial. A discussion on fitness landscape is included, which paves the way for Kauffman's NK model to analyse to performance of genetic algorithms and their operators. Future work which may extend from this could include using the NK model to analyse the performance of the inversion operator.

Download the Full Paper

Iterated Prisoner's Dilemma: A review

Technical Report No. : NUIG-IT-260601
By : C. O'Riordan

Much debate has centered on the nature of social dilemmas. In environmental issues, trade-wars, negotiations between countries and social interactions, there are many scenarios where involved players may choose altruistic, cooperative action or an alternative selfish behaviour. Games, particularly the prisoner's dilemma and variations have been used to model and analyse such scenarios. In this paper, an overview of research in this domain is presented. The iterated prisoner's dilemma (IPD) is discussed with attention payed to evolution of strategies in this environment, the effect of noise in the IPD and finally N-player versions.

Download the Full Paper

A Virtual Laboratory Model for Coordinating Future Astrobiological Research

Technical Report No. : NUIG-IT-310701
By : A. Golden, A. Shearer, R. Butler, J. Browne, A. Penny and J. Nuth

Future multidisciplinary problems in European Astrobiological research will require a large effort across a broad spectrum of work activities. These activities are likely to involve resources and personnel distributed amongst the various ESA member states. Current and evolving informatic technologies could, by means of centralising and distributing information through a `Virtual Laboratory' (VL) environment, greatly enhance such work. We propose to implement such a VL environment to coordinate research into prebiotic chemistry in the early Solar Nebula. This will involve facilities in both a number of ESA member states and the US, and personnel similarly distributed. We outline preliminary aspects of the VL that require particular attention, focus on critical informatic requirements, and assess the overall outcome of such a programme of research. We emphasise the considerable efforts required to provide a working `real time' interactive environment for the participants.

Download the Full Paper

Organizational context and the IS implementation process: an integrated research framework

Technical Report No. : NUIG-IT-270801
By : G.J. Lyons and T. Geraghty

Good IS managers have an intuitive grasp of the complex organizational and project dynamics required for successful implementation of large information systems. Years of hard-earned experience have taught them that IS implementation requires simultaneous attention to the political and structural contexts of the host organization as well as to the execution of a well planned IS project. Yet, much of the research on IS implementation has focused on narrowly bounded, individual dimensions of this multi-faceted real-world phenomenon. The work described here attempts to integrate these various perspectives into a robust research strategy, and evaluates its application in a comparative case study. Broadly, research on information systems implementation can be classified into two distinct groups: Factors Studies and Process Studies. These offer valuable academic insight into separate aspects of IS implementation concerned, respectively, with the underlying factors affecting IS success or failure and the dynamics of the implementation project. However, the practitioner must grapple with both of these dimensions simultaneously, attempting to achieve a pragmatic balance between business objectives, organizational and political dynamics and the implementation process itself. This paper presents an integrated methodology for studying systems implementation within its organizational and business contexts. The approach uses a number of models and techniques to formulate an holistic view of the implementation process, to enable its examination and subsequent interpretation in an organizational framework. Development of the methodology involved the selection and adaptation of key elements from cognate areas such as IS investment analysis and success criteria, techno-structural approaches to organization development, and IS implementation research. The methodology was successfully tested in a comparative case study on the implementation of a major information system in two separate business units of an electrical utility. This integrated approach provides an holistic IS implementation framework for practitioners and helps to knit together much valuable earlier work into a coherent research strategy.

Download the Full Paper

Bridging the Gap between Desirable Web-based Education Systems and Current Technical Solutions

Technical Report No. : NUIG-IT-220901
By : P. Byrne, S. Flynn, J. Griffith and C. O' Riordan

The increased use of the web as a means of disseminating material seems a dominant trend in third level institutions. This modern development is the latest in a long trend in harnessing technology to mimic the traditional classroom model. However, systems to-date do not deal with many of the issues recognised in pedagogical research and instead merely present information via a new medium. Techniques from AI seem to offer the potential to address some of the oft overlooked ev idence from the pedagogy research community. This paper discuss some topics from the arena of education research and their incorporation into automated educational systems. A review of many of the existing approaches is presented. The paper also outlines ideas from AI which offer hope in bridging the gap between the desirable requirements of web-based education systems and current systems.

Download the Full Paper

An investigation into the use of synchronised time to improve VoIP service

Technical Report No. : NUIG-IT-240901
By : H. Melvin and L. Murphy

This paper presents an overview of work in progress relating to the use of synchronised time in Voice over IP (VoIP) networks. One of the principal limitations of conventional IP networks is the best-effort service that they provide. In recent years, the Internet Engineering Task Force (IETF) have created the multimedia data and control architecture which incorporates protocols such as the Realtime Transport Protocol (RTP), RTP Control Protocol (RTCP) and Real Time Streaming Protocol (RTSP). Whilst these protocols assist in delivery of multimedia traffic, the lack of bounds on end-to-end delay still presents significant problems for interactive applications such as VoIP. The ITU-T recommendation G.114 specifies that a round-trip-delay of 300ms should not be exceeded and the Plain Old Telephone System (POTS) generally does much better than this. Along with RTP and RTCP, various approaches have been proposed and taken to improve the performance of VoIP networks from an interactive viewpoint. These range from sender-based codec measures to network-based policies that differentiate between traffic flows and finally to various buffer schemes at the receiver that balance overall delay with receiver packet loss due to late arrival. This paper summarises much of this work but focuses on receiver-based measures. RTP timestamps are principally used to enable receiver hosts to monitor and react to the inter-packet delay variance (jitter). RTCP packet timestamps further enable sender hosts to periodically determine round-trip-times (RTT). Such calculations do not require that end host clocks are synchronised. In this paper, a scheme is described whereby synchronised time is available to hosts in a VoIP session. As Global Positioning System (GPS) receivers become increasingly cost-effective and always-on Internet connections more prevalent, VoIP hosts can avail of synchronised time either via GPS or Network Time Protocol (NTP). This time synchronisation enables each host to know precisely the end-to-end delays on a packet-by-packet basis. This is a significant improvement on the present situation whereby sender hosts can periodically determine round-trip-times and thus estimate one-way-delays. Most adaptive receiver buffer schemes do not consider such estimates and are designed to minimise delay at the expense of tolerable packet loss due to late arrival and some distortion of inter-talkspurt silence periods. In many instances, the performance of private IP networks and the localised public Internet will operate well within the bounds of G.114. In such situations, a fixed end-to-end delay within the G.114 bounds yet marginally above the actual end-to-end performance would result in a tolerable delay with no late packet loss and no distortion of silence periods. With synchronised time, supporting protocols and receiver-based intelligence, such a system can be implemented.

Download the Full Paper

An Architecture for Information Retrieval from Distributed Heterogeneous Information Sources

Technical Report No. : NUIG-IT-250901
By : . Ryan and C. O' Riordan

This paper provides an overview of the initial design of a system to provide accurate responses to users' queries against a set of information repositories. We present an overview of techniques from traditional information retrieval and filtering, techniques from the more specialised field of distributed information retrieval (paying particular attention to the problems of source selection and result fusion) and, finally, we discuss the initial design for our system.

Download the Full Paper
 

• Disclaimer
• Privacy
• Copyright
• Accessibility