PAA
2007
13 years 10 months ago
2007
Feature selection methods are often used to determine a small set of informative features that guarantee good classification results. Such procedures usually consist of two compon...
HCI
2007
14 years 10 days ago
2007
This study shows that in a qualitative study a small sample size is sufficient to gain interesting results and show differences between the procurement of IT services in different ...
ACII
2005
Springer
14 years 26 days ago
2005
Springer
The null space N(St) of total scatter matrix St contains no useful information for pattern classification. So, discarding the null space N(St) results in dimensionality reduction ...
MICCAI
2000
Springer
14 years 2 months ago
2000
Springer
We present a novel approach to statistical shape analysis of anatomical structures based on small sample size learning techniques. The high complexity of shape models used in medic...
MCS
2009
Springer
14 years 3 months ago
2009
Springer
The meta-learner MLR (Multi-response Linear Regression) has been proposed as a trainable combiner for fusing heterogeneous baselevel classifiers. Although it has interesting prope...
ICPR
2002
IEEE
14 years 12 months ago
2002
IEEE
The small sample size problem is often encountered in pattern recognition. It results in the singularity of the within-class scatter matrix Sw in Linear Discriminant Analysis (LDA...
ICIP
2004
IEEE
15 years 15 days ago
2004
IEEE
It is well-known that the applicability of Linear Discriminant Analysis (LDA) to high-dimensional pattern classification tasks such as face recognition (FR) often suffers from the...
CVPR
2004
IEEE
15 years 27 days ago
2004
IEEE
Invariant features or operators are often used to shield the recognition process from the effect of "nuisance" parameters, such as rotations, foreshortening, or illumina...