Game Development Reference
In-Depth Information
on video sequences. In their work, they introduce the concept of high-level/low-
level analysis. In their approach, the high-level analysis structure takes as input
the FAP produced by the low-level analysis tool and, by means of an HMM
classifier, detects the facial expression on the frame.
Fuzzy systems
Fuzzy systems are an alternative to traditional notions of set membership and
logic. The notion central to fuzzy systems is that true values (in fuzzy logic) or
membership values (in fuzzy sets) are indicated by a value on the range [0.0, 1.0],
with 0.0 representing the absolute Falseness and 1.0 representing absolute
Truth. This is a new approach to the binary set 0 (False) — 1 (True) used by
classical logic. Fuzzy systems try to gather mathematical tools to represent
natural language, where the concepts of True and False are too extreme and
intermediate or more vague interpretations are needed.
Apart from the basic operations among sets, fuzzy systems permit the definition
of “hedges,” or modifiers of fuzzy values. These operations are provided in an
effort to maintain close ties to natural language, and to allow for the generation
of fuzzy statements through mathematical calculations. As such, the initial
definition of hedges and operations upon them is quite a subjective process and
may vary from one application to another. Hedges mathematically model
concepts such as “very,” “somewhat,” “sort of,” and so on.
In many applications fuzzy systems appear as a complement to the image
processing involved; they help in the decision-making process needed to evaluate
results from analyzed images. Huntsberger, Rose and Ramaka (1998) have
developed a face processing system called Fuzzy-Face that combines wavelet
pre-processing of input with a fuzzy self-organizing feature map algorithm. The
wavelet-derived face space is partitioned into fuzzy sets, which are character-
ized by face exemplars and memberships values to those exemplars. The most
interesting properties for face motion analysis which this system presents are
that it improves the training stage because it uses relatively few training epochs
and that it generalizes to face images that are acquired under different lighting
conditions. Fellenz et al. (2000) propose a framework for the processing of face
image sequences and speech, using different dynamic techniques to extract
appropriate features for emotion recognition. The features are used by a hybrid
classification procedure, employing neural network techniques and fuzzy logic,
to accumulate the evidence for the presence of an emotional facial expression
and the speaker's voice.
Search WWH ::
Custom Search