Multiple Scale and Multimodal Data and Information Fusion in Human
- Involving: Fionn Murtagh and Dimitri Zervas, Royal
Holloway, University of London; and Glyn Humphreys and Roberta Roberts,
University of Birmingham
- Funded by: Innovative Biological Imaging and Signal Analysis, BBSRC,
Reports and Presentations
- "Feature Detection Aided by
Comodulated Noise", presentation, 29 March 2007.
- "Haptic Feature Detection Through Comodulated
Noise", report, 11 February 2007, 12 pp.
We consider the detection of faint signal, very much
below the noise level. The series of signals is
presented to subjects for their
assessment, by haptic means, of the presence of the feature in one or other
of the two signals presented. We develop an approach based on visualization
of the dependencies between the signal, possibly containing the feature, and
another noise signal.
- Report, 13 May 2007, 18 pp.
R.D. Roberts, G.W. Humphreys and A.M. Wing, "Temporal
constraints on ineractions across kinaesthetic
subject responses are evaluated
relative to correct outcomes (i.e.,
experiment design conditions). Here we include this aspect
in the analysis but also
carry out two "Turing test type evaluations". The latter are
based on two algorithms, which generally (but not always) outperform
the test subjects.
It is not our purpose to simply outperform the subjects: our algorithms
benefit strongly from prior information. Our algorithms are used
to baseline the subject responses. A multivariate data analysis not only
account the correct outcomes, the subject performance, and our two
also takes into account differentiation within the subject population.
Correlation Effect", report, 13 July 2007, 25 pp.
We describe the requirements for the modelling of the correlation
effect, the detailed algorithm, the program used, and sample output produced.
Contents: (1) Introduction, (2) Algorithm for Modelling Correlation Effect,
(3) File coreffect.h, (4) File coreffect.cpp, (5) Sample Output.
Background for this Work and Its Subsequent Development
Next, further development.
- Faint signal detection in environmental data: section 3.2.5,
"From Fuzzy Coding to Possibility Theory", in Murtagh (2005) -- see below.
- Faint signal detection in financial data: section 4.5, "Financial
Modeling and Forecasting", with subsections: "Brownian Motion",
"Granularity of Coding", "Fingerprinting of Price Movements", in
Murtagh (2005) -- see below.
Faint signal detection in time-varying, correlated data:
- F. Murtagh, A. Ganz, S. McKie, J. Mothe and K. Englmeier, "Tag Clouds for Displaying Semantics: The Case of Filmscripts", Information Visualization Journal, 9, 253-262, 2010.
- F. Murtagh, M. Spagat and J.A. Restrepo, "Ultrametric Wavelet Regression of Multivariate Time Series: Application to Colombian Conflict Analysis", IEEE Transactions on Systems, Man, and Cybernetics--Part A: Systems and Humans, 41, 254-263, 2011.
- Eye Mouse,
human computer interaction based on eye gaze.
- Multiresolution Image and
Signal Processing, software and books online
(Image Processing and Data Analysis: The Multiscale Approach,
JL Starck, F Murtagh and A Bijaoui, Cambridge University Press, 1998;
Astronomical Image and Data Analysis,
JL Starck and F Murtagh, Springer, 2002. 2nd Edition, 2006).
Multivariate Data Analysis Software and Resources, and
Software Accompanying: Correspondence Analysis and Data Coding
with R and Java, F Murtagh, 2005.
Last update: 2012 February 7, by f m u r t a g h * a t * a c m * s p o t *
o r g