13.01.2015 Views

Face Detection and Modeling for Recognition - Biometrics Research ...

Face Detection and Modeling for Recognition - Biometrics Research ...

Face Detection and Modeling for Recognition - Biometrics Research ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

(a)<br />

(b)<br />

(c)<br />

Figure 3.2. Skin detection: (a) a yellow-biased face image; (b) a lighting compensated<br />

image; (c) skin regions of (a) shown in white; (d) skin regions of (b).<br />

(d)<br />

method. Note that the yellow bias color in Fig. 3.2(a) has been removed, as shown in<br />

Fig. 3.2(b). The effect of lighting compensation on detected skin regions can be seen<br />

by comparing Figs. 3.2(c) <strong>and</strong> 3.2(d). With lighting compensation, our algorithm<br />

detects fewer non-face areas <strong>and</strong> more skin-tone facial areas. Note that the variations<br />

in skin color among different racial groups, reflection characteristics of human<br />

skin <strong>and</strong> its surrounding objects (including clothing), <strong>and</strong> camera characteristics will<br />

all affect the appearance of skin color <strong>and</strong> hence the per<strong>for</strong>mance of an automatic<br />

face detection algorithm. There<strong>for</strong>e, if models of the lighting source <strong>and</strong> cameras are<br />

available, additional lighting correction should be made to remove color bias.<br />

<strong>Modeling</strong> skin color requires choosing an appropriate color space <strong>and</strong> identifying<br />

a cluster associated with skin color in this space.<br />

It has been observed that the<br />

62

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!