Professional Documents
Culture Documents
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 20
Abstract— Caricature is a pictorial description of a person or subject in a summarizing way using exaggeration of the most distinguish features and over-
simplification of the common features in order to make that subject ‘unique’ and to preserve the recognizable likeness of the subject. Facial caricature gen-
erator is developed to assist the user in producing facial caricature automatically or semi-automatically. It is derived from the rapid advance in computer
graphics and computer vision as well as introduced as a part of non-photorealistic rendering technologies. Recently, facial caricature generator becomes
particularly interesting research topic due to the advantageous features of privacy, security, simplification, amusement and their rampant emergent real-
world application such as in magazine, digital entertainment, Internet and mobile application. This paper reviews the uses of caricature in variety of applica-
tions, theories and rules in the art of drawing caricature, how these theories are simulated in the development of caricature generation system and the current
research trend in this field. There are two main categories of facial caricature generator based on their input data type: human centered approach and image
centered approach. It also briefly explains the general process of generating caricature. The state of the art techniques in generating caricature are described
in detail by classifying it into four approaches: interactive, regularity-based, learning-based and predefined database of caricature illustration. Expressive
caricature is also introduced which is evolved from the neutral caricature. This paper also discusses relevant issues, problems and several promising direc-
tions for future research.
Index Terms—caricature, exaggeration, face image, facial feature extraction, computer graphics and vision.
—————————— ——————————
1 INTRODUCTION
in generating facial caricature from face image. Section 7 of caricature in internet and mobile phone application [5],
briefly addresses the expressive caricature. In Section 8, [6], [7], [8], [9]. J. Liu et al. [5] and R. Q. Zhou and J. X. Liu
we discuss several issues, problems and the promising [7] changed the people’s real photograph captured by
direction for facial caricature generator research. Lastly, mobile camera to black and white caricature and then it
section 9 summarizes and concludes this review paper was used to synthesize the multimedia animation mes-
sage and transferred it to the mobile terminal as an enter-
2 APPLICATION OF CARICATURE tainment. M. Lyons et al. [8] used caricature of human
face in creation personalized avatar system. J. Liu et al. [6]
2.1 Caricature in Entertainment and Political Magazine proposed a prototype system for online chatting in which
Caricature has been widely used in our daily life since the caricature was used for people to express their emotion or
past few decades. It always can be seen in magazine or to show their identity to the others they communicating
newspaper for various different purposes. Caricatures of with M. Sato et al. [9] used caricature with a small set of
movie stars are often displayed in entertainment maga- characteristic points for non-verbal communication in
zines. It is used for entertainment through the combined mobile phone.
use of fantasy and humor. Caricature is also used for ex-
pression of social and political perspective such as criti- 2.3 Caricature in Face Recognition
cizing intolerance, injustice, political corruption and so- Use of caricature for face recognition and perception has
cial evils using humor and satire. Caricatures of politi- been extensively studied in cognitive psychology, visual
cians are commonly met in editorial cartoons. Political perception, computer vision, and pattern recognition ar-
caricatures generally are thought provoking and strive to ea. Caricature is often used in facial recognition especially
educate the viewer about a current issue. A typical news- to recognize the criminal faces in police investigation or
paper article has a great many words to deliver infor- to identify subjects in forensic process. The victims or
mation and ideas whereas a political caricature can ad- witnesses are often asked to describe the facial features of
dress issues to ground level and expose the reader to the criminal and then, an artist will draw the caricature
sharp ridicule by reducing an entire article down to sim- according to the description. Caricatures are generally
ple pictures. Political caricatures have been proved as a perceived as better likenesses of a person than the veridi-
powerful vehicle for swaying public opinion and criticiz- cal images [10]. It is because the exaggeration of the dis-
ing or praising political figures [3]. tinctive feature in caricature emphasizes the features of
the face and enhances the recognition. The distinctive
2.2 Caricature in Internet and Mobile Application features are important in face recognition and faces with
With the emergence of Internet and mobile technology, a unusual features are better recognized than more typical
caricature has been used for social communication and faces [11]. Caricature is not only used during recognized
entertainment over web or mobile phone. The user can faces but also used to increase the amount of memorable
protect their identity and real image from other users for information of the face encoded for later recognition.
security purposes but still allows the basic facial gestures People often recognize faces from their own races better
to be recognizable. Caricature is very applicable to be than those from other races [12]. J. Rodriguez et al. [12]
used in network environment because of their simplifica- used caricature to reduce other-race effect in facial recog-
tion in representing a person, which can reduce the com- nition and to help individuals focus their attention to the
putational burden and the bandwidth requirement, com- features that are useful for recognition. G. Rhodes et al.
pared to the real photo image. Caricature is used as a dis- [13] found that line drawing caricatures can be better rec-
play picture to create a memorable impression of a par- ognized than the veridical image but photographic carica-
ticular user in visual messenger, visual chat room, bulle- ture did not. C. Frowd et al. [10, 14], and J. Smitaveja et al.
tin board and forum. It is also used as an avatar in virtual [15] employed caricature in face recognition system using
community such as virtual museum, virtual classroom, in different approaches.
interactive movie and multiple user role-playing commu-
nity in virtual game. Avatar is a graphical representation 2.4 Caricature in Information Visualization and
of a person in virtual environment. According to K. L. Education
Nowak and C. Rauh [4], the people preferred an avatar Basic concept of caricature that exaggerated the character-
that is matched to their own gender, type (human) and istic features of the subject and simplifies the overall
with other characteristics that are similar to their own. structure is adopted in information visualization for bet-
Thus, use of facial caricature of the people as an avatar ter understanding about the subject. It provides a power-
might be a good preference since the caricature can pro- ful metaphor for illustrative visualization. It is called as
vide users with valuable information about their partners caricaturistic visualization [16]. Caricature is also used to
and can reduce uncertainty in their interaction. Caricature visualize the differentiation of subjects or concepts that
is also used as an avatar in a blog, website, profile page hard to be noticed or to ease the user in understanding
on the social networking and dating sites. Other than that, the difference of subjects. The potential applications are
caricature also can be used in e-business card, e-greeting such as quality control, comparative biology, case-based
card, e-logo or as a perfect gift idea for many occasions. education for medical students and deformation surveil-
Many research have been carried out concerning the use lance. Caricature is also used to enhance learning espe-
JOURNAL OF COMPUTING, VOLUME 3, ISSUE 5, MAY 2011, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 22
cially for the complex learning. It helps the cognitive sys- steps, observation and exaggeration.
tem to pick up distinctive features of the learned material Observation is the most important things to do in
and minimizes any perceptual or representational confu- drawing caricature. The likeness can be lost very easily if
sion [17]. the observation is not carefully accomplished. Caricaturist
need to understand what they see in the individual face
2.5 Caricature in User Interface Agent that makes him recognizable, observe which of the facial
Caricature is also used as an interface agent to assist a features are larger, smaller, sharper, and rounder than
user in accomplishing daily computer tasks such as sort- most people’s, identify the salient facial features that
ing email, filtering information and scheduling meetings. make him unique and extract his personal interpretation.
It can help the user interact with the agent and under- Person’s distinguishing characteristic can be determined
stand the agent’s behavior or characteristic through their by comparing his/ her facial features with the reference.
external trait [18]. Caricature could also make a computer According to the psychologist [13], human beings have a
more human-like, engaging, entertaining, approachable, “mean face” recorded in their brain, which is an average
and understandable to the user, establish relationships of faces they encounter in life. This “mean face” acts as a
with users, and make the users feel more comfortable reference. The distinguishable facial feature is the feature
with computers [19]. It also creates strong implications in that larger or smaller or sharper or rounder than average.
user interaction with the system, which makes the user However, a caricaturist considers the difference not only
more likely to be cooperative. from the average face but also their surrounding features
of the same subject. L. Redman [22] employed the princi-
3 THEORIES AND RULES IN THE ART OF DRAWING ple of relativity to define the distinctive facial features.
CARICATURE
Relativity in the arts can be thought of as having two
parts: the relationship of things to others of their own
3.1 Elements of Caricature kind and the relationship of things to their surrounding
There are three essential elements that must be presented and abutting elements. The first part gives us the infor-
in caricature [20],[21]. The first element is Likeness. All mation about the size or shape of the facial features com-
good caricature must have a good likeness of their sub- pare to others such as the eye is larger than others where-
jects. If the caricature cannot represent who it is supposed as, the second part tells about the distance or aspect ratio
to be, then it is not a caricature. The second element is between the facial features such as distance between eyes
exaggeration. All caricature must have some form of ex- or ratio of the mouth width to the nose width. L. Redman
aggeration or a departure from the exact representation of [22] used in-betweener as a frame of reference for deter-
the subject features. If there is no exaggeration, what we mining how to exaggerate the subjects’ features. He exag-
have is a portrait but not a caricature. The degree of exag- gerates any features of the subject that is different from
geration employed by each individual caricaturist varies in-betweener. T. Richmond [20] defined the relationship as
enormously. Exaggeration is the essence of a caricature the distances between the five simple shapes, their size
and not a distortion. Exaggeration is the overemphasis of relative to one another, and angle they are at in relation-
truth, while distortion is a complete denial of truth [22] ship to the center axis of the face. The five simple shapes
for instance if the mouth of a person is prominent, make it are left eye, right eye, nose, mouth, and face shape. He
enormous and not make it tiny. However, there are ex- used classical portrait proportion as a point of reference
traordinary creative caricaturists who succeed in dis- to observe where the subject's face might differ and de-
torting some part of facial feature without losing their cide what relationships to change and how much to
likeness. The last element is statement or subjective im- change them. The width of an eye is used as the primary
pression. Caricature must have something to say about frame of reference. It is difficult to observe those faces in
the subject. It might be something to do with the situation which the unique features are not obvious. In order to
the subject is drawn in, their personality through expres- make observation easy, simplify the face into five basic
sion or body language or visual fun of some aspect or shapes by eliminating the details.
their image. The second step is exaggeration. The way of the
3.2 Process of drawing caricature caricaturists draw and exaggerate a caricature is depend
Drawing a caricature is an inborn talent exits only in a on their style. Most of the artists start drawing a carica-
few peoples [23]. This inborn talent embedded in their ture with the simplest form or outline of the subject after
subconscious mind and hard to be explained. However, they have determined the unique features or the relation-
these talented people still have to be taught because ship to be exaggerated from their observation. The simple
sometimes this talent may go unnoticed or undiscovered shape is easier to draw, control and exaggerate to the de-
until later of life. Drawing caricature needs eye-hand co- sired shape than ones with a lot of complex element. The
ordination, which uses the eyes to pay attention to the caricaturist always creates several little drawing of differ-
subject and the hands to draw the subject based on the ent exaggerated facial features and then picks the best one
information received through the eyes. In addition, we [1]. After that, the detail of facial features will be added to
need our brain to find the most prominent facial features create the likeness. Which facial features are the most im-
that make the person different from everyone else [21]. portant part and should be looked and drawn first is var-
The process of drawing caricature involves two basic
JOURNAL OF COMPUTING, VOLUME 3, ISSUE 5, MAY 2011, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 23
ies to different artist. K. Bjorndahl [1] , and Shafali [21] related to the linguistic terms. The increase and the de-
consider that the eyes is the most important feature on the crease of the parameter value are defined according to the
face, whereas T. Richmond [20] believes that the face Linguistic term. Linguistic terms used to describe the fa-
shape is the most important part and then eye and nose. cial features are such as “big”, “small”, “round”, “thin”
The exaggeration is the manipulation of the relationship and “thick”. Impression terms are such as “pretty”, “se-
of the features to one another in term of their distance, vere”, “honesty”, “active”[28], “cheerful”, “quiet”, “ma-
size and angle and not the features themselves taken in- ture”, “youthful” [30]. The general procedure for generat-
dividually. Even the eyes is bigger than other peoples, it ing facial caricature from word description is shown in
cannot be simply drawn in bigger size. The rest of the face Fig 1. The user will input words that describe their de-
should be taken into account. Every change to one feature sired facial caricature image. The words have been de-
will affect to the others features or called as action and fined in the fuzzy set. Some calculation will be done ac-
reaction. For example, if the eyes are moved farther apart, cording to the inputted term to get the features parameter
the nose will move closer to the eyes. If the chin is values. The correspondence image in the database will be
changed to become bigger, the top of a head will be located [30] or the average face that is used as an initial
smaller. This technique will create higher impact exag- face will be transformed to the desired facial caricature
geration. Other than that, basic knowledge of drawing, according to the parameter value [27], [29].
anatomical knowledge and experience in drawing a hu-
man portrait are very helpful in drawing caricature. Input
4.1 Human Centered Approach Finally, facial caricature with obtained parameter values
is presented to the user. If the output is not fit to the de-
A facial caricature is generated based on the user’s verbal
sired facial caricature image, some modification can be
description of facial features and the impressions about
done also using the linguistic term. S. Iwashita et al. [27],
the target face. This approach attempts to imitate how the
S. Iwashita et al. [29], and T. Onisawa and Y. Hirasawa
suspect’s face is drawn based on verbal description given
[30] focused on how to seek the desired image of the face
by the witness in police investigation. There have been
so that the output really reflect the face image in the us-
some studies on generating facial caricature from the us-
er’s mind while J. Nishino et al. [28] concentrated on the
er’s linguistic description [27], [28], [29], [30]. These stud-
linguistic acquisition and emphasizing on the structure of
ies are using fuzzy theory to express the meaning of the
the fuzzy set. The problems that always encountered in
linguistic term or word and to represent the parameter
these studies are the ambiguity meaning of the linguistic
value of each facial feature. These parameter values are
term, the display of the facial features is not really similar
JOURNAL OF COMPUTING, VOLUME 3, ISSUE 5, MAY 2011, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 24
stroy the likeness of the subject and more interesting but ing to the SRM.
it has high computation and storage complexity whereas
the lower number of facial feature points can save the
computation and storage cost but the result is less inter-
esting and limited to certain exaggeration. Table 1 below
shows the landmark points used in some research works.
Table 1
Number of Landmark Points Used in Some Research Works
Basic Research works
features Fig 3 Canon model [47]
[43] [44] [45] [46] [32] [23] [9]
eyes 16 16 16 40 44 8 12
eyebrows 20 20 20 14 16 8 16
nose 12 12 12 25 22 5 3
mouth 17 20 20 24 18 6 10
plied than the mesh morphing method but it takes more algorithm, which utilize the interactive morphing tool to
time to compute [50], [51]. The most common problem in generate caricatures as shown in Fig 6. He starts with an
morphing is to specify corresponding feature points of extremely simple template, which represents the essential
the images that can provide reasonable result and reduce features by a few numbers of lines. Only one feature is
computational complexity [52] and to morph only specific exaggerated at a time. He uses try and error method to
parts of the face image while holding other parts constant find the distinctive feature to be exaggerated. If that ex-
[50]. The metamorphosis techniques that implemented by aggeration did not create a likeness, go in the opposite
the previous works in generating caricature are mesh direction. If this too did not give the required result, re-
morphing [23], [35], [36], feature-based morphing [32], turn the feature to its original position. The user must
[53] or known as field morphing, Thin plate spline warp- correctly identify the distinctive features. It is difficult to
ing [43], and Delaunay Triangulation warping [54]. Fur- do for the ordinary user and take a long time in doing try
thermore, Arruda et al. [42] used deformable mesh tem- and error process to get the desired result. The user is also
plate, M. Obaid et al. [40] employed quadratic defor- might exposed to the risk of producing unrecognizable
mation model and E. Akleman et al. [53] used implicit caricature. E. Akleman et al. [53] further came up with a
free form deformation. C. C. Tseng et al. [44], B. Gooch et new deformation technique that uses simplical complex
al. [48] and Z. Mo et al. [55] did not state the specific tech- to generate caricature. He uses simplices (point and trian-
nique of morphing. gle) as a basic deformation primitive. Triangles can pro-
The final caricatures produced by the existing vide shear transformation, which cannot be supported by
works come in a variety of styles such as photographic line. The deformations are defined by a set of simplex
caricature [23], [35], [36], [54], [55], [56] as shown in Fig pair. Blending function is used to interpolate the transla-
5(a), sketch caricature [43], [44], [55] in Fig 5(b), black- tion vector, which define the source to target mapping.
white illustration caricature [41], [48] and hand-drawn- This system can intuitively and interactively produce an
like caricature as shown in Fig 5(c) [32]. Some of other extreme caricature but requires experience to meaningful-
works generate outline caricatures as shown in Fig 5(d) ly specify source and target simplices. All these works
by connecting the exaggerated facial feature points using require expert knowledge or skilled user input, which
lines [31],[34], [57], [58], [59], piecewise cubic Hermite limits their applicability for every-day use [48].
interpolation [37], Bezier curves [39], and B-Spline curves
[9], [38],[46].
(a)
(a) (b)
(b)
Fig 6 (a) original face image and its simple template [53] (b) caricature
image and its simple template [53]
(c) (d)
Fig 5 (a) Photographic caricature [55] (b). sketch caricature [44] (c) hand- 6.2 Regularity-based Approach
drawn like caricature [32] (d) outline caricature [34] Regularity based approach is summarizing the rules of
creating caricature and simulated by computer to gener-
ate caricature automatically or semi-automatically. The
6 DIFFERENT APPROACHES IN GENERATING basic rule in drawing caricature is exaggerating the dif-
ference from the reference face. Average face is widely
FACIAL CARICATURE FROM INPUT FACE IMAGE
agreed among the caricaturists and the researchers to be
used as a reference face. S. E. Brennan [31] used an aver-
6.1 Interactive Approach
age face derived from a collection of face images in data-
In this approach, the distinctive facial features and the
base as a reference to find distinctive features. The posi-
exaggeration rate are determined interactively by the us-
tions of 165 feature points on the input face image corre-
er. The user can drag the points of the facial features to
sponding to the reference points in the average face are
enlarge or reduce the size in order to generate the desired
marked manually with mouse click. These points are then
caricature. E. Akleman [60] came up with a very simple
JOURNAL OF COMPUTING, VOLUME 3, ISSUE 5, MAY 2011, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 27
compared with the positions of similar points on the av- pression to the user. T. Kamimura and Y. W. Chen [38], C.
erage face in order to determine the extent of deviation. C. Tseng [44], S. J. Gibson et al.[54], and X. Guangzhe et
This deviation would further be exaggerated by moving al. [59] employed Principle Component Analysis (PCA) to
away the feature points from the average to yield a carica- find the major principle components of features. The pa-
ture as shown in Fig 7. The rate of exaggeration was de- rameter of exaggeration rate is determined empirically by
fined interactively by hand control. The facial caricature is performing orthogonal expansion to the eigenvector [59].
represented in line drawing that connects all these feature Varying the coefficient of eigenvector would change the
points. This system still requires skill user input to control face shape. However, Z. Mo et al.[55] has a strong persua-
the feature points in order to generate an interesting cari- sion that EDFM method may not produce the best carica-
cature. tures. The distinctiveness of a displaced feature not only
depends on its distance from the mean, but also its vari-
ance for example even eye width and mouth width devi-
ate from the mean in equal distance, the eye will be more
distinctive compared to the mouth if the eye width has
much less widely distributed. Instead of using mean face
as a reference to find the distinctive features, some works
using standard face model as a reference. B. Gooch et al.
[48] proposed an interactive method in deforming the
black and white facial illustration to produce caricatures
(a) (b) (c)
which rely much less on the presence of well-trained us-
Fig 7 (a) input face image (b) line representation of input face image [31] ers. He uses a norm face grid (Fig 4) as a reference and a
(c) caricature image [31]
facial feature grid (FFG) for an input face image which
consist a rectangle frame, four vertical line that marking
This rule is also used by [23] ,[26] ,[32], [39], [41], [55],
the inner and outer corners of the eyes and three horizon-
[56], [61]. P. Y. Chiang et al. [32] carried out with a differ-
tal lines mark the position of the eyes, the tip of the nose,
ent result as shown in Fig 8 but still using an average face
and the mouth as shown in Fig 9(a). Scaling the vectors
as a reference face. An artist’s work is used as a source
between corresponding vertices in the norm face grid and
image to be morphed according to the exaggerated points
the user-set grid does the exaggeration. To allow more
of original face image to produce a caricature. The notion
expressive freedom, vertices on the left and right of the
of “exaggerating the difference from the mean” (EDFM) is
grid can be manipulated individually but not the internal
also used by [34],[57],[58],[62],[63]. This rule can be de-
features in order to produce a caricature as shown in Fig
fined in formula as
9(b) and Fig 9(c).
Q P b P S (2)
where P is input images, Q is generated caricature, S is
the average face, b is the rate of exaggeration and (P-S) is
the difference between input image and the average face.
caricaturist. C. C. Tseng [44], S. J. Gibson[54], and X. exaggerations and not only linear exaggeration, which
Guangzhe [59] determined the exaggeration rate empiri- scale the features with a factor as provided in most of the
cally by eigenvector and by a simple linear model [32]. existing caricature generation systems. They proposed a
neural network based caricature generation. Feed for-
Table 3
ward back propagation [23] and cascade correlation [65]
Regularity-based approach are adopted to capture the relationship between the dis-
Relativity tinctive facial features and the changes from original face
Reference
face Relationship to image to caricature. The distinctive facial features are de-
Relationship to others
the surrounding
termined by comparing face image to the average face.
[26], [31], [34], [38],
Average This system cannot generate caricature exactly the same
[39], [41], [55],[56], [44], [39]
face as the artist drawing but still can be accepted as satisfac-
[57],[58], [61], [62],[63]
tory result as shown in Fig 10. J. Liu et al.[35] and J. Liu et
Standard
[47] , [48], [49] al. [36] learn the style of general artist by using hand
face model
drawn caricatures that are created by many artists over
the world while K. H. Lai et al.[23], L. Liang et al.[43] and
N. S. Rupesh [65] learn an individual artist style and the
6.3 Learning-based Approach caricatures are drawn by a particular artist. K. H. Lai et
al.[23], L. Liang et al.[43], and N. S. Rupesh [65] known as
This approach is always used to capture the artist style or
supervised learning whereas Liu et al.[35] and J. Liu et al.
to simulate the caricaturist creativity that are difficult to
[36] as semi-supervised learning.
codify algorithmically such as the style of the particular
artist exaggerate the facial features, the style of stroke and
the style of sketch. The artist’s product will be learned
using machine learning technique or artificial intelligent
methods. This approach requires extensive training data-
bases containing pairs of original face image and its cor-
responding caricatured face image drawn by caricaturist.
L. Liang et al.[43] proposed the caricature generating sys-
tem based on example. He analyses 92 face image-
caricature pairs and classifies into 28 prototypes using (a) (b) (c)
Partial Least Square (PLS). Each prototype contains sam- Fig 10: (a) original face image (b) hand-drawn caricature created by an
ples with similar direction of exaggeration correspond to artist (c) the resulting caricature [23]
some facial features. Any input of face image would be
classified into one prototype. The facial features to be ex- 6.4 Predefined Database of Caricature Illustration
aggerated and the rate of exaggeration are estimated us-
The distinctive features and exaggeration rate are not de-
ing local linear model (least square). The exaggeration
termined explicitly in this approach. Illustration carica-
direction and the selected facial features determined by
tures correspond to the original facial parts need to be
this method are limited and the distinctive facial features
defined beforehand [46], [66], [67]. This approach is also
selected by the artist may cover different prototypes. J.
known as component-based [68]. An input face image
Liu et al. [35] proposed a mapping learning approach to
will be extracted and caricature illustration closest to the
generate facial caricature. He employs Principle Compo-
extraction result would be selected from a database. T. W.
nent Analysis (PCA) to obtain the principle component of
Pai and C. Y. Yang [46] concentrated on classifying and
the facial features and uses Support Vector Regression
drawing methodology in order to ease the matching pro-
(SVR) to learn the mapping model in principle compo-
cess between an input face image and the pre-generated
nent space. A new input face photograph would be pro-
caricature in facial bank. The resulting caricature generat-
jected into the principle component space and the vector
ed by T. W. Pai and C. Y. Yang [46] shown in Fig 11. The
that represents the position of the target caricature could
problem of this approach is to obtain all the required cari-
be predicted using the mapping model. This vector needs
cature illustration in advance. If the correct caricature
to be transformed from the principle component space to
illustration is not available in the database, the facial part
high-level feature space to get the caricature’s shape. Both
of an input face image might be replaced with unsuitable
of the works [35]and [43] used a linear methods to map
one and resulted in producing unrecognizable caricature.
the original face image to its corresponding facial carica-
Sometime, even the individual parts of the input face im-
ture whereas J. Liu et al. [36] further came up with non-
age are matched flawlessly, the resulting caricature bear
linear mapping model. He employs semi-supervised
very little resemble to the person [68].
manifold regularization learning to reduce dimensionali-
ty of the caricature and to build regressive model in order
to predict the caricature pattern from the given original
face image. K. H. Lai [23] and N. S. Rupesh [65] also be-
lieve that generating facial caricature involves non-linear
JOURNAL OF COMPUTING, VOLUME 3, ISSUE 5, MAY 2011, ISSN 2151-9617
HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/
WWW.JOURNALOFCOMPUTING.ORG 29
Fig 11 (a) original face image (b) matched face contour (c) caricature image
[46] Fig 12 (a) input face image (b) natural expression (c) smile expression (d)
surprise expression [40]
the facial feature points in another face image. They used esting subject research.
their experience and judgment to select the position of the
facial feature points, which is prone to human mistake, 8.7. Non-photorealistic Rendering Techniques
and take a lot of time. The questions here are how to find
the corresponding feature locations in different faces? This technique such as pen-and-ink, brush, hatch, paint-
Does the location of facial feature points in one face is ing etc need to be improved and enhanced to render the
accurately correspond to the facial feature points in an- facial caricature texture and the style of artist drawing
other face? need to be captured and combine with the existing non-
photorealistic rendering technique in order to let the cari-
8.4. Learning-based Approach cature approaching the artist’s product.
312 , 2000,.
[63] T. Yamaguchi, M. Tominaga, and H. Koshimizu, ʺInteractive
facial caricaturing system based on Eye Camera,ʺ in Proceed‐
ings of SPIE ‐ The International Society for Optical Engineer‐
ing. vol. 5132, pp. 390‐399, 2003.
[64] Y. Q. Xu, H. Y. Shum, M. Cohen, L. Liang, and H. Zhong, ʺcar‐
icature exaggeration,ʺ United States Patents Application Pub‐
lication, vol. US 2005/0212821 A1, sept 29, 2005.
[65] N. S. Rupesh, H. L. Ka, A. E. Eran, and W. H. C. Paul, ʺUse of
Neural Networks in Automatic Caricature Generation: An
Approach Based on Drawing Style Capture,ʺ in Lecture note
in computer science : Pattern Recognition and Image Analysis,
vol. 3523/2005: Springer Berlin, 2005.
[66] C. Y. Yang, T. W. Pai, Y. S. Hsiao, and C. H. Chiang, ʺFace
shape classification for 2D caricature generation,ʺ in Proceed‐
ings of Workshop on Consumer Electronics, 2002.
[67] F. Min, J. L. Suo, S. C. Zhu, and N. Sang, ʺAn Automatic Por‐
trait System Based on And‐Or Graph Representation ʺ in Lec‐
ture Note in Science Computer, 2007.
[68] W. H. Liao and Chien An Lai, ʺAutomatic Generation of Cari‐
catures with Multiple Expressions Using Transformative Ap‐
proach,ʺ in Arts and Technology 2009, LNICST 2010, 30 ed, pp.
263‐271, 2010.
[69] A. J. Calder, D. Rowland, A. W. Young, I. Nimmo‐Smith, J.
Keane, and D. I. Perrett, ʺCaricaturing facial expressions,ʺ in
Cognition, vol. 76, pp. 105‐146, 2000.
[70] O. Mohammad, M. Ramakrishnan, and B. Mark, ʺGenerating
and rendering expressive caricatures,ʺ in ACM SIGGRAPH
2010 Posters Los Angeles, California: ACM, 2010.
Suriati bte Sadimon received her BSc in Science, Computer with Educa-
tion from Universiti Teknologi Malaysia and her MSc in Distributed Mul-
timedia System from University of Leeds, United Kingdom. She is current-
ly persuing her Phd study at Universiti Teknologi Malaysia. Her area of
interest includes computer graphics, face image processing and machine
learning.
Mohd Shahrizal Sunar received his BSc in Computer Science from Uni-
versiti Teknologi Malaysia, his Msc in Computer Graphics and Virtual
Environment from University of Hull, United Kingdom and his Phd in
Computer Graphics from Universiti Kebangsaan Malaysia. Currently, he is
a senior lecturer and a head of department of Computer Graphics and Mul-
timedia, Universiti Teknologi Malaysia. His area of interest includes com-
puter graphic, virtual reality, game programming and image processing.