|
Pattern Matching & Visualisation
The predominant goal of Pattern Recognition is the classification of
objects into a number of categories or classes. Patterns can arise from
a wide range of phenomena, e.g. images, signal waveforms, microarray
data, financial market movements and so on. To a great extent, the
ability to perceive, measure and interpret patterns goes hand-in-hand
with the ability to form a mental or physical image of them, so that
visualisation (or the visualisation tool in computer terms) is often
referred to in the same breath. While attractive, the image is only part
of the story, however, and the modelling techniques are the core.
Pattern Recognition has a long history, which is rooted in statistical
theory, but in the latter part of the 20th century, computational
advances enormously increased the demand for practical applications of
pattern recognition and these have in turn stimulated new theoretical
developments. For example, Pattern Recognition is an integral part in
most machine intelligence systems built for decision-making. Similarly,
machine vision is an area of key importance and is used for a variety of
tasks, such as finding "defects" in automated quality processes,
determining "hot spots" in medical diagnosis, interpretation of gesture
and speech and so on. Areas of key interest to group members include
recognition of hand gestures from sign language for the Deaf such as ISL
(Irish Sign Language), Spatio-temporal Gesture recognition and
Human-Computer Natural Interfaces. In the work on ISL, for example,
images are blurred at different stages in order to make a hierarchical
database of shapes and a Hidden Markov Model uses the database
information to extract the shapes for recognition. In HCI, and in order
to interact with computers with no physical connection, a gesture
recognition system is required.
|
|
|
Additionally, and very recently, the successful mapping of the Human
Genome has increased the already-rapid rate of advance in molecular
biology and has spawned a number of new inter-disciplinary areas. While
widely-used misnomers, such as Bioinformatics, suggest that most of the
work is in database and information retrieval, the impact on
inter-disciplinary science is, in fact, far greater. The field of
microarray data analysis, for example, has been around for less than
ten years and is the focus of large and growing efforts of statisticians,
biologists and others. Large-scale, high throughput assays may involve
parallel collection, systematic and random error analysis and of course
classification and interpretation of gene expression data. Pattern
recognition and visualisation have a major role to play in new biological
experimentation methods.
Biocomputation and Models, Microarray, GA etc. - M. Crane, H.J. Ruskin, A. Barat
ISL, Gesture Recognition and MV Section - A. Sutherland, H. Wu
Researchers: Martin Crane, Heather J. Ruskin, Alistair Sutherland, Wu Hai, Thomas Coogan, George Awad
|
|
