FAST ADAPTABLE SKIN COLOUR

DETECTION IN RGB SPACE

Martin Tosas

and Steven Mills

School of Computer Science and IT, University of Nottingham, Nottingham, United Kingdom

Geospatial Research Center (NZ) Ltd, University of Canterbury, Christchurch, New Zealand

Keywords: Skin Colour Detection, Linear Container, Human Computer Interaction.

Abstract: This paper presents a skin colour classifier that uses a linear container in order to confine a volume of the

RGB space where skin colour is likely to appear. The container can be adapted, using a single training

image, to maximize the detection of a particular skin tonality. The classifier has minimum storage

requirements, it is very fast to evaluate, and despite operating in the RGB space, provides equivalent

illumination (brightness) independence to that of classifiers that work in the rg-plane. The performance of

the proposed classifier is evaluated and compared with other classifiers. Finally, conclusions are drawn.

1 INTRODUCTION

We propose a skin colour classifier named the

Linear Container (LC) classifier. The classifier uses

four decision planes in order to confine a volume of

the RGB space where skin colours are likely to

appear. The features of the LC classifier are: capable

of being tuned to a particular skin tonality; rapid

evaluation; minimal storage requirements; and

resistance to illumination (brightness) changes

equivalent to that of classifiers that work in

normalised RGB. The classifier requires a tuning

stage in which a single image, with marked skin and

background areas, is analysed resulting in a new

linear container configuration. The classifier is

originally intended to detect skin colour in tracking

applications, such as the ones found in Human

Computer Interaction (HCI) systems; however, other

uses are conceivable.

The paper is organized as follows: Section 2

outlines previous work on skin colour detection, and

situates the LC classifier in relation to other works;

Section 3 introduces the LC classifier, and describes

its tuning procedure; Section 4 presents performance

results; Section 5 shows the behaviour of the LC

classifier when tuned at various resolutions; Section

6 studies some HCI usability factors; Finally,

Section 7 gives some conclusions and directions for

future work.

2 PREVIOUS WORK ON SKIN

COLOUR DETECTION

Skin colour provides an important source of

information for computer vision systems that

monitor people. The skin colour cue is widely used

in face detection and recognition systems, various

types of surveillance, vision-based biometric

systems, and vision-based HCI systems. All these

areas of application use skin colour to track, locate

and interpret people, with relatively efficient, fast,

low-level, methods.

The goal of skin colour detection is to build a

decision rule that can discriminate between the skin

and non-skin colour pixels of an image. Because of

the importance of skin colour detection there have

been numerous approaches to solve this task. The

various approaches can be grouped into the

following four categories: non-parametric skin

distribution modelling, parametric skin distribution

modelling, explicitly defined skin region modelling,

and dynamic skin colour modelling (Vezhnevets et

al., 2003).

Non-parametric skin distribution modelling uses

training data to estimate a skin colour distribution.

This estimation process is sometimes referred to as

the construction of a Skin Probability Map (SPM)

(Jones and Rehg, 1999; Brand and Mason, 2000;

Gomez, 2002) assigning a probability value to each

Tosas M. and Mills S. (2007).

FAST ADAPTABLE SKIN COLOUR DETECTION IN RGB SPACE.

In Proceedings of the Second International Conference on Computer Vision Theory and Applications, pages 3-10

DOI: 10.5220/0002055800030010

 SciTePress

point of a discretised colour space. A SPM can be

implemented by a colour histogram, and such

approaches normally use the chrominance plane of

some colour space in order to offer resistance to

illumination changes (Chen et al., 1995; Schumeyer

and Barner, 1998; Jones and Regh, 1999; Zarit et al.,

1999). SPMs can use a Bayes classification rule in

order to improve their performance, in this case two

colour histograms are required; one for the

probability of skin colour, and another for the

probability of non-skin colour (Jones and Regh,

1999; Zarit et al., 1999; Chai and Bouzerdoum,

2000). The main disadvantages of SPMs are the high

storage requirements and the fact that their

performance directly depends on the

representativeness of the training images.

Parametric skin distribution modelling can

represent skin colour in a more compact form.

Common examples of parametric modelling model a

skin colour distribution using a single Gaussian

(Ahlberg, 1999; Menser and Wien, 2000; Terrillon

et al., 2000) or a mixture of Gaussians (Jones and

Regh, 1999; Yang and Ahuja, 1999; Terrillon et al.,

2000). Expectation Maximization (EM) algorithms

are used on training data to find the model

parameters that produce the best fit. The goodness of

fit, and therefore the performance of the model,

depends on the shape of the chosen model and the

chosen colour space. This performance dependency

with the colour space is stronger in the case of

parametric modelling than it is in the case of non-

parametric modelling (Brand and Mason, 2000; Lee

and Yoo, 2002).

Another way to build a skin colour classifier is to

define explicitly, through a number of rules, the

boundaries of a skin cluster in some colour space;

this is called explicitly defined region modelling.

The obvious advantage of this method is its

computational simplicity, which has attracted many

researchers (Fleck et al., 1996; Ahlberg, 1999; Jorda

et al., 1999; Peer et al., 2003), as it leads to the

construction of a very rapid classifier. However in

order to achieve high recognition rates both a

suitable colour space and adequate decision rules

need to be found empirically. Gomez and Morales

(2002) proposed a method that can build a set of

rules automatically by using machine learning

algorithms on training data. They reported results

comparable to the Bayes SPM classifier in RGB

space for their data set.

Finally, we have dynamic skin colour modelling.

This category of skin modelling methods is designed

for skin detection during tracking. Skin detection in

this category is different from static image analysis

in a number of aspects. First, in principle, the skin

models in this category can be less general – i.e

tuned for a specific person, camera, or lighting.

Second, an initialisation stage is possible, when the

skin region of interest is segmented from the

background by a different classifier or manually; this

makes possible to obtain a skin classification model

that is optimal for the given conditions. Finally, this

category of skin models can be able to update

themselves in order to match changes in lighting

conditions. Some of the methods in this category use

Gaussian distribution adaptation (Yang and Ahuja,

1998), or dynamic histograms (Soriano et al., 2000;

Stern and Efros, 2002). In (Soriano et al., 2000) a

skin locus, in rg space, is constructed beforehand

from training data. Then, during tracking, their

dynamic skin colour histogram is updated with

pixels from the bounding box of the target, provided

these pixels belong to the skin locus. This makes the

dynamic histogram less likely to adapt to colour

distributions other than that of skin.

The proposed LC classifier belongs to the last

two categories. The classifier is implemented using

rules similar to the rules of the explicitly defined

skin region models; however, these rules are

parameterised in order that they can be tuned to

specific conditions, during an initialisation stage.

The parameters of the LC classifier can also be

recalculated rapidly in order to adapt to changing

illumination conditions.

3 LINEAR CONTAINER

CLASSIFIER

Normalised RGB is a popular colourspace because

of its simple normalisation procedure, and its

diminished dependence with brightness (Yang and

Ahuja, 1998; Zarit et al., 1999; Lee and Yoo, 2002;

Stern and Efros, 2002; Peer et al., 2003). The

projection from RGB (3D space) to normalised RGB

(2D space) corresponds with a cone in the original

3D RGB space, in that each point in the rg-plane

corresponds to a 3D line of colour values in the

original RGB space. These lines meet at (0, 0, 0),

and points along the lines correspond to scaling of

white illumination. Therefore, a skin colour cluster

in the rg-plane corresponds to a cone-like cluster in

RGB space. This is illustrated in Figure 1.

The proposed LC classifier uses a polyhedral

cone, constructed from four decision planes, in order

to model the cone-like region in RGB space that

VISAPP 2007 - International Conference on Computer Vision Theory and Applications

(a)

(b)

Figure 1: (a) Skin colour cluster, in the rg-plane, from a single sample. (b) The rg-plane skin colour cluster projected to

RGB space; each point in the rg-plane becomes a line in RGB space.

Figure 2: (a) Horizontal decision planes. (b) Vertical decision planes. (c) Tuning heuristic.

results from the projection of a skin colour cluster in

the rg-plane to the RGB space. The LC classifier

performs pixel-based segmentation. If an RGB value

is inside the polyhedral cone volume, it is classified

as skin; if the RGB value is outside the polyhedral

cone volume then it is classified as non-skin. The

definition of the four decision planes is:

Ghmin G BRmin R B BGhmax G BRmax R⋅+ ⋅< < ⋅+ ⋅

(1)

where BGhmin and BRmin parameterise the lower

"horizontal" plane, and BGhmax and BRmax

parameterise the higher horizontal plane. The

horizontal planes are illustrated in Figure 2(a). These

two planes confine a volume between them by

constraining the values that B can take in relation to

R and G. This volume is further constrained by two

"vertical" planes:

Gvmin B GRmin R G BGvmax B GRmax R⋅+ ⋅< < ⋅+ ⋅

(2)

where BGvmin and GRmin parameterise the left

vertical plane, and BGvmax and GRmax

parameterise the right vertical plane. The vertical

planes confine a volume between them by

constraining the values that G can take in relation to

R and B. The vertical planes are illustrated in

Figure2(b). As the RGB values that are close to the

origin carry too little colour information, we truncate

the apex of the polyhedral cone with one additional

rule:

min R

. If a colour value satisfies these three

rules, then it is inside the polyhedral cone, and

therefore classified as skin colour.

The LC classifier can be tuned for a specific

person, camera, or lighting conditions, in an

initialisation step. For this, an initialisation image is

needed. This initialisation image is composed of two

approximately complementary masks; one mask

delimits the target skin colour area, we call this

mask SkinMask; and the other mask comprises areas

where we do not expect to find skin colour, we call

this mask BackgroundMask. Figure 3 shows an

initialisation image segmented by the two masks.

The BackgroundMask can be tailored in order to

avoid areas of skin colour in addition to those

included in SkinMask, for example, Figure 3(b)

avoids the subject's wrist. The two masks can be

generated manually, or automatically by a tracking

system such as (Tosas and Li, 2007).

The tuning procedure uses a heuristic method by

which the parameters of the decision planes are

changed in sequence.

(a)

(b)

(c)

FAST ADAPTABLE SKIN COLOUR DETECTION IN RGB SPACE

(a) (b)

Figure 3: Initialisation image segmented by SkinMask (a)

and BackgroundMask (b).

Each time a parameter is changed, the fitness of the

LC classifier, to the detection of skin colour in the

SkinMask and to the rejection of skin colour in the

BackgroundMask, is measured using the following

equation:

# #

kin in SkinMask skin in BackgroundMask

fitness TI

ize of SkinMask size of BackgroundMask

=×−

(3)

where TI (Target Importance) is used to control the

importance of the target skin colour area in the

fitness. In the experiments of the following sections

TI = 2 so as to give double importance to detecting

skin on the SkinMask than to avoid detecting skin on

the BackgroundMask. This parameter allows the

classifier to be tuned to favour true positives or

negatives.

The heuristic search, by which the parameters of

the decision planes are changed, is illustrated in

Figure 2(c). This figure shows a section view of the

RGB cube, corresponding to the B-G-plane with

maximum R. Lines 1, 2, 3 and 4 are the intersections

of the four decision planes with the section view.

Starting from some priori values, the search varies

BRmin, then BRmax, GRmin, and finally GRmax;

first, reducing their values, then increasing their

values, and measuring the fitness (Equation 3) at

each step. The values that produce the best fitness

are finally selected. Note that the slope of the

decision planes remains unchanged in this heuristic.

4 PERFORMANCE RESULTS

The LC classifier is tested on video sequences of

subjects with four different skin tonalities:

Mediterranean, white Caucasian, black African, and

Chinese. The target skin colour area is the subject's

hand. The subjects hold their hand open in front of

the camera, and move the hand towards and away

from the camera. An overhead lamp affects the

illumination of the subjects' hand. When the

subject's hand is closer to the camera, the hand is

under a shadow and looks darker. When the subject's

hand is further away from the camera, the hand is

under the lamp and looks brighter. The classifier is

initialised once, at the beginning of each video

sequence.

The skin colour detection performance is

calculated for each video sequence, using a ground

truth. The ground truth consists of two masks, which

have been manually generated for every fifth frame

of the four video sequences. The ground truth

considers the subject's hand as the target area for

skin colour detection. This area is segmented using

the SkinTruth mask, Figure4(b). The background is

segmented using the BackgroundTruth mask, Figure

4(c). Note that the BackgroundTruth mask is not the

complement of the SkinTruth mask. The

BackgroundTruth mask avoids the target skin colour

area, the subject's hand, and any other skin colour

areas in the image; therefore, for each measurement

frame, there will be some areas which will not take

part in the counting; these areas correspond to the

subject's face and arms. Both masks are tested for

skin colour. Skin colour pixels found in the

SkinTruth mask constitute true-positives. Non-skin

colour pixels found in the BackgroundTruth mask

constitute true-negatives. In order to compare

detection results between frames the true-positives

and true-negatives are normalised to the size of

SkinTruth and BackgroundTruth masks respectively.

Normalised true-positives are referred to as NTN,

and normalised true-negatives are referred to as

NTP.

Figure 4: (a) Original frame. (b) SkinTruth mask. (c)

BackgroundTruth mask.

We use an rg skin colour histogram classifier as a

comparison reference with the LC classifier. The rg

histogram used for comparison is constructed in an

initialisation step at the beginning of each sequence

from the pixels in SkinMask and its size is 64x64

bins. A pixel is classified as skin colour if its

corresponding bin in the rg histogram is bigger than

a threshold.

The choice of the threshold affects the detection rate

of the rg histogram. In general, if the threshold

increases, NTN tends to be higher, but NTP tends to

be lower; if the threshold decreases, NTP tends to be

VISAPP 2007 - International Conference on Computer Vision Theory and Applications

higher, but NTN tends to be lower. For the tested

video sequences a threshold of 25 produced the best

results.

Frame 90

Frame 110

Figure 5: Mediterranean subject test. Top row: Original

frames. Middle row: rg histogram classifier. Bottom row:

LC classifier.

Frame 15

Frame 85

Figure 7: White Caucasian subject test. Top row: Original

frames. Middle row: rg histogram classifier. Bottom row:

LC classifier.

Figure 5 shows the results for the Mediterranean

subject. The figure presents plots of the NTP and

NTN against the frame number, and two example

frames showing the skin colour classification for a

best detection case and a worst detection case. The

top row shows the original frames before detecting

the skin colour areas. The middle row corresponds to

the rg histogram, and the bottom row corresponds to

the LC classifier. The results of both classifiers are

similar, but the LC classifier consistently exhibits

slightly larger NTP and NTN than the rg histogram

classifier, all along the video sequence. Following

the same format as Figure 5, Figures 6, 7 and 8 show

the results for the other three ethnic skin tonalities.

In all cases, the LC classifier exhibited the same or

larger NTP and NTN than the rg histogram

classifier.

Frame 10

Frame 70

Figure 6: Black African subject test. Top row: Original

frames. Middle row: rg histogram classifier. Bottom row:

LC classifier.

Frame 25

Frame 85

Figure 8: Chinese subject test. Top row: Original frames.

Middle row: rg histogram classifier. Bottom row: LC

classifier.

Videos showing the skin colour detection tests are

available at:

www.cs.nott.ac.uk/~mtb/

research/SkinColour.html

An experiment comparing the computational speed

of the LC classifier, against other classifiers, was

also carried out. The experiment consists in

measuring the time it takes for a classifier to check

all the pixels in a 640x480 frame. The experiment

is repeated for 100 frames of a video sequence

containing skin colour regions, and the times used in

each frame are averaged.

FAST ADAPTABLE SKIN COLOUR DETECTION IN RGB SPACE

Table 1: Execution time results.

Average time per frame

Speed-up of the RGB LC classifier with

respect to the other classifiers

RGB LC classifier 0.0090 secs

rg LC classifier 0.0147 secs x1.62

rg LC classifier with lookup table 0.0107 secs x1.17

rg histogram 0.0235 secs x2.59

rg histogram with lookup table 0.0204 secs x2.25

bare RGB histogram 0.0022 secs x0.24

The experiment was carried out in an AMD Athlon

3500+, 1GB of RAM. The results are shown in

Table 1.

The RGB LC classifier uses an extra rule to

avoid dark pixels, the other classifiers do not use this

rule. The rg LC classifier is the 2D equivalent to the

proposed RGB LC classifier. It works in the rg-plane

by using 4 decision lines instead of 4 decision

planes. The skin detection performance of this

classifier is equivalent to the RGB LC classifier. The

equations in the rg LC classifier are simpler than

those of the RGB LC classifier; however, the former

is slower because it has to normalise each pixel from

RGB to rg. The use of lookup table containing all

the possible normalisations can speed up the

normalisation procedure. But, even when using a

lookup table, the RGB LC classifier is x1.172126

times faster than its rg LC equivalent. And the rg LC

classifier is faster than the rg histogram classifier

(both with and without lookup table).

A bare RGB histogram classifier is used as a

comparison measure for the fastest skin colour

classifier (yet the most sensitive to changes in

illumination), being x4.16 times faster than the LC

classifier. However, in practice, the reduced storage

requirements of the LC classifier may result, when

plugging it into certain algorithms, in faster

execution times due to its better locality of

reference. This is illustrated in a practical

application of the LC skin colour classifier. The LC

classifier is used in the measurement function of the

particle-filter based hand contour tracking algorithm

described in (Tosas and Li, 2007). In such a tracking

algorithm, most of the computation is expended in

the measurement function (profiling shows that

more than 60% of the application's time is spent in

the measurement function). The average time spent

in the processing of a tracking time-step is

calculated as the average of the times spent in each

of 100 time-steps of tracking. When exchanging the

LC classifier for a 32x32x32 bins RGB histogram

(in the same machine as the previous experiment)

the speed up in the processing of a time-step is only

x1.2 times faster (as opposed to the x4.16 times

faster suggested in the previous experiment).

5 TUNING AT VARIOUS

RESOLUTIONS

So far, the LC classifier has been tuned using an

initialisation image of the same size as the video

sequence in which it was tested, 640x480 pixels. It

was observed that the tuning of the LC classifier on

a decimated version of the initialisation image,

results in little degradation of the classifier's

detection performance on the non-decimated video

sequence. This is because the result of the tuning is

more dependent upon the range of colours of the

pixels in the initialisation masks than upon the

number of pixels. This fact allows us to speed-up the

tuning procedure, because the amount of data to be

dealt with is reduced. The speed-up of the tuning

procedure as a result of using a decimated

initialisation image instead of using a non-decimated

initialisation image is: x4 for a 320x240 resolution,

x16 for 160x120, x64 for 80x60, x256 for 40x30,

and x1024 for a 20x15 resolution. Figure 9 shows

the NTP of the LC classifier, on the video sequence

of the Mediterranean subject, for various resolutions

Figure 9: NTP when tuning at various resolutions.

VISAPP 2007 - International Conference on Computer Vision Theory and Applications

of the initialisation image. The NTN are not shown

as they remain at almost 1 for the six resolutions.

Notice that the NTP for an initialisation image of

320x240 is virtually the same as the NTP for an

initialisation image of 640x240.

The speed-up resulting from the use of

decimated initialisation images becomes very

important in HCI applications because it allows the

tuning (and potentially periodical retuning) of the

LC classifier with only a small impact in the HCI

system responsiveness.

6 HCI USABILITY FACTORS

The tuning stage in the experiments of the previous

sections was idealised, in that no background

colours appear in the SkinMask, and no skin colour

appeared in BackgroundMask. If the LC classifier is

used in a HCI system, which could generate the

initialisation masks automatically from a tracking

subsystem, it is possible that background appears in

SkinMask, and skin colour appears in

BackgroundMask. In this section we study the

robustness of the LC classifier against non-ideal

tuning conditions.

The detection performance of the LC classifier is

calculated, once more, for the video sequence of the

Mediterranean subject. This time, the tuning is

repeated for a misaligned SkinMask and

BackgroundMask. In each repetition SkinMask only

contains a percentage of the target's skin colour area.

The skin that is not in the SkinMask is in the

BackgroundMask, this affects the final configuration

of the LC parameters found during the tuning stage.

Figure 10 shows the NTP for four percentages of

skin colour in SkinMask. The NTN is not shown as

it is almost unaffected in all the four cases. We can

see that the degradation in NTP for a 50% skin in

SkinMask is small; and even when the amount of

skin in SkinMask is as small as 25%, the NTP along

the whole sequence may still be useful for some

applications. However, the model parameters found

during the tuning stage, depend on the colours

appearing in each initialisation mask; hence,

different results are possible even when SkinMask

contains the same amount of skin. This is illustrated

in Figure 11, where the tuning of the LC classifier

using two SkinMasks with the same percentage of

skin inside the mask, produce different detection

performances.

(a)

(b)

(c)

(d)

Figure 10: Chart, NTP for four percentages of skin in

SkinMask. (a) SkinMask containing 100% skin, (b) 75%

skin, (c) 50% skin, and (d) 25% skin.

25%A

25%B

Figure 11: NTP for two SkinMask containing 25% skin

colour.

7 CONCLUSIONS

We have presented the Linear Container (LC) skin

colour classifier. This classifier constitutes a

contribution to dynamic skin colour modelling

methods. Its detection performance compares well

with an rg histogram classifier, resulting in equal or

better detection rates, when using a single training

image. Two remarkable qualities of this classifier

are its evaluation speed, and its low storage

requirements. The four rules that define the decision

planes, and an extra rule to avoid dark pixels, can be

rapidly evaluated, resulting in a x2.24 speed-up with

respect to a simple rg histogram classifier. In

practice, the reduced storage requirements of the LC

classifier may result, when plugging it into certain

algorithms, in even faster execution times due to its

better locality of reference. This can prove to be an

advantage in embedded systems. On the other hand,

despite the LC classifier operates in the RGB space,

its resistance to illumination changes is equivalent to

that of a classifier that operates in the rg-plane. The

detection performance of the LC classifier is not

greatly impaired when the tuning is performed in a

decimated initialisation image, but the execution

FAST ADAPTABLE SKIN COLOUR DETECTION IN RGB SPACE

time of the tuning is notably reduced. The LC

classifier also proved to be robust to non-ideal

initialisations, in which skin colour appears in

BackgroundMask, and background appears in

SkinMask.

A subject of further work is the tuning stage.

Different heuristics or maximisation procedures

could produce better detection results. On the other

hand, the LC model itself could be changed. Linear

containers are fast to evaluate, but other type of

containers, could produce a better fit of the skin

colour cluster through scaling of white illumination.

Sets of rules such as the ones proposed in (Gomez

and Morales, 2002) could give better detection

results, although a tuning procedure for these rules

may be more complex.

REFERENCES

Ahlberg, J., 1999. A system for face localization and facial

feature extraction. Tech. Rep. LiTH-ISY-R-2172,

Linkoping University.

Brand, J. and Mason, J., 2000. A comparative assessment

of three approaches to pixel level human skin-

detection. In Proc. of the ICPR, vol. 1, 1056-1059.

Chai, D., And Bouzerdoum, A., 2000. A Bayesian

approach to skin color classification in ycbcr color

space. In Proc. of IEEE Region Ten Conference

(TENCON’2000), vol. 2, 421-424.

Chen, Q., Wu, H., and Yachida, M., 1995. Face detection

by fuzzy pattern matching. In Proc. of the ICCV, 591-

597.

Fleck, M., Forsyth, D. A., and Bregler, C., 1996. Finding

naked people. In Proc. of the ECCV, vol. 2, 592-602.

Gomez, G., 2002. On selecting colour components for skin

detection. In Proc. of the ICPR, vol. 2, 961-964.

Gomez, G., and Morales, E., 2002. Automatic feature

construction and a simple rule induction algorithm for

skin detection. In Proc. of the ICML Workshop on

Machine Learning in Computer Vision, 31-38.

Jones, M. J. and Rehg, J. M., 1999. Statistical color

models with application to skin detection. In Proc. of

the CVPR ’99, vol. 1, 274-280.

Jorda, L., Perrone, M., Costeira, J., and Santos-Victor, J.,

1999. Active face and feature tracking. In Proc. of the

International Conference on Image Analysis and

Processing, 572-577.

Lee, J. Y., and Yoo, S. I., 2002. An elliptical boundary

model for skin color detection. In Proc. of the 2002

International Conference on Imaging Science,

Systems, and Technology.

Menser, B., and Wien, M., 2000. Segmentation and

tracking of facial regions in color image sequences. In

Proc. Visual Communications and Image Processing,

SPIE, 731-740.

Peer, P., Kovac, J., and Solina, F., 2003. Human skin

colour clustering for face detection. In International

Conference on Computer as a Tool, EUROCON, The

IEEE region 8, vol 2, 144-148.

Vezhnevets, V., Sazonov V., Andreeva A., 2003. A

Survey on Pixel-Based Skin Color Detection

Techniques. In Proc. Graphicon, pp. 85-92, Moscow,

Russia.

Schumeyer, R., And Barner, K., 1998. A color-based

classifier for region identification in video. In

Proc.Visual Communications and Image Processing,

SPIE, vol 3309, 189-200.

Soriano, M., Martinkauppi, B., Huovinen, S., and

Laaksonen, M., 2000. Using the skin locus to cope

with changing illumination conditions in color-based

face tracking. In Proc. of the IEEE Nordic Signal

Processing Symposium, pp. 383-386.

Stern, H., and Efros, B., 2002. Adaptive color space

switching for face tracking in multi-colored lighting

environments. In Proc. of the International

Conference on Automatic Face and Gesture

Recognition, 249-255.

Terrillon, J. C., Shirazi, M. N., Fukamachi, H., and

Akamatsu, S., 2000. Comparative performance of

different skin chrominance models and chrominance

spaces for the automatic detection of human faces in

color images. In Proc. of the International Conference

on Face and Gesture Recognition, 54-61.

Tosas, M., and Li, B., 2007. Tracking Tree-Structured

Articulated Objects Using Particle Interpolation.

Accepted for publication in the proceedings of

CGIM2007.

Yang, M., and Ahuja, N., 1998. Detecting human faces in

color images. In Proc. of ICIP, vol. 1, 127-130.

Yang, M. H., and Ahuja, N., 1999. Gaussian mixture

model for human skin color and its applications in

image and video databases. In Proc. of the SPIE:

Conf. On Storage and Retrieval for Image and Video

Databases, vol. 3656, 458-466.

Zarit, B. D., Super, B. J. and Quek, F. K. H., 1999.

Comparison of five color models in skin pixel

classification. In ICCV’99 Int’l Workshop on

recognition, analysis and tracking of faces and

gestures in Real-Time systems, 58- 63.

VISAPP 2007 - International Conference on Computer Vision Theory and Applications