face detection

14
Face detection Tuck N.

Upload: vmk

Post on 02-Dec-2014

996 views

Category:

Technology


7 download

DESCRIPTION

 

TRANSCRIPT

Page 1: Face detection

Face detection

Tuck N.

Page 2: Face detection

IDEF0 decomposition

Page 3: Face detection

Face detection’s approaches

• Contour images analysis• Intensity and colors analisys• Adaptive classifier

Page 4: Face detection

Contour images analysis

Page 5: Face detection

Face detection’s approaches

Page 6: Face detection

Face detection’s approaches

Skin color’s cluster

Page 7: Face detection

Adaptive classifier

Page 8: Face detection

Adaptive classifier

Page 9: Face detection

AdaBoost

AdaBoost is an algorithm for constructing a “strong” classifier as linear combination

f(x) =αtht(x)

of “simple” “weak” classifiers ht(x) : X → {−1,+1}.

Page 10: Face detection

Rapid Object Detection using a Boosted Cascade of Simple Features

Integral Images

Page 11: Face detection

Algorithm workflow

Page 12: Face detection

Classifier’s algorithm workflow

Page 13: Face detection

OWL model

We can use OWL model to store some data, used in our application. For example, I’ve stored maximum and minimum area of faces, that should be detected. According to w3c standarts:

<owl:Restriction> <owl:onProperty rdf:resource="#minWidth" /> <owl:minCardinality rdf:datatype="&xsd;nonNegativeInteger">20</owl:minCardinality></owl:Restriction><owl:Restriction> <owl:onProperty rdf:resource="#maxWidth" /> <owl:maxCardinality rdf:datatype="&xsd;nonNegativeInteger">300</owl:maxCardinality></owl:Restriction>

Page 14: Face detection

OWL and C# interaction

We can use standart XmlDocumentReader to parse OWL model. Also we can use LINQ to parse this model.