Smile Detection by Boosting Pixel Differences

Report
Caifeng Shan, Member, IEEE
IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 21, NO. 1,
JANUARY 2012

INTRODUCTION

METHOD

EXPERIMENTS

INTRODUCTION

METHOD

EXPERIMENTS



Most of the existing works have been focused on
analyzing a set of prototypic emotional facial
expressions
Using the data collected by asking subjects to pose
deliberately these expressions
In this paper, we focus on smile detection in face
images captured in real-world scenarios

INTRODUCTION

METHOD

EXPERIMENTS
BOOSTING PIXEL DIFFERENCES
 S. Baluja
and H. A. Rowley, “Boosting set
identification performance,”Int. J. Comput. Vis.,
vol. 71, no. 1, pp. 111–119, 2007
 Baluja
introduced to use the relationship
between two pixels’ intensities as features.
 they
used five types of pixel comparison
operators (and their inverses):
 The
binary result of each comparison, which is
represented numerically as 1 or 0, is used as
the feature. Thus, for an image of pixels, there
are
or 3312000 pixelcomparison features
 Instead
of utilizing the above comparison
operators, we propose to use the intensity
difference between two pixels as a simple
feature
 For
an image of 24*24 pixels, there are
or 331200 features extracted
AdaBoost ( Adaptive Boosting )
AdaBoost learns a small number of weak classifiers
whose performance is just better than random
guessing and boosts them iteratively into a strong
classifier of higher accuracy
the weak classifier
consists of feature
(i.e.,
the intensity difference),threshold
, and parity
indicating the direction of the inequality sign as follows:

INTRODUCTION

METHOD

EXPERIMENTS
Data
 Database
: GENKI4K
consists of 4000 images (2162 “smile” and 1828
“nonsmile”)
 In
our experiments, the images were converted
to grayscale
 the
faces were normalized to reach a canonical
face of 48*48 pixels
Data
Illumination Normalization
 Histogram equalization (HE)

Single-scale retinex (SSR)

Discrete cosine transform (DCT)

LBP

Tan–Triggs
Illumination Normalization
Boosting Pixel Intensity Differences
Average of (left) all smile faces and (right) all nonsmile faces
Impact of Pose Variation
Impact of Pose Variation
Thank you

similar documents