一篇图像分割的外文论文，可供大家学习

image

1星需积分: 0 326 浏览量更新于2023-05-21 评论 2 收藏 4.95MB PDF 举报

身份认证购VIP最低享 7 折!

领优惠券(最高得80元）

资源详情

资源评论

资源推荐

Neurocomputing 292 (2018) 1–27

Contents lists available at ScienceDirect

Neurocomputing

journal homepage: www.elsevier.com/locate/neucom

Segmentation of images by color features: A survey

Farid Garcia-Lamont

∗

, Jair Cervantes

, Asdrúbal López

, Lisbeth Rodriguez

Universidad Autónoma del Estado de México, Centro Universitario UAEM Texcoco, Av. Jardín Zumpango s/n, Fraccionamiento El Tejocote, CP 56259,

Texcoco-Estado de México, México

Universidad Autónoma del Estado de México, Centro Universitario UAEM Zumpango, Camino viejo a Jilotzingo continuación Calle Rayón, CP 55600,

Zumpango-Estado de México, México

Instituto Tecnológico de Orizaba, División de Investigación y Estudios de Posgrado, Av. Oriente 9, 852. Col. Emiliano Zapata, CP 94320, Orizaba-Veracruz,

México

a r t i c l e i n f o

Article history:

Received 13 August 2017

Revised 30 October 2017

Accepted 8 January 2018

Available online 7 March 2018

Communicated by Chennai Guest Editor

Keywords:

Color spaces

Image segmentation

Quantitative evaluation

a b s t r a c t

Image segmentation is an important stage for object recognition. Many methods have been proposed in

the last few years for grayscale and color images. In this paper, we present a deep review of the state

of the art on color image segmentation methods; through this paper, we explain the techniques based

on edge detection, thresholding, histogram-thresholding, region, feature clustering and neural networks.

Because color spaces play a key role in the methods reviewed, we also explain in detail the most com-

monly color spaces to represent and process colors. In addition, we present some important applications

that use the methods of image segmentation reviewed. Finally, a set of metrics frequently used to evalu-

ate quantitatively the segmented images is shown.

Introduction

Image segmentation is one of the most important object recog-

nition stages for artiﬁcial vision systems. Image segmentation is

deﬁned as the union of sets that contains the pixels coordinates

with an speciﬁc feature; in other words, let I

= ∪

i =1

be the seg-

mented image, such that ∩

i =1

= ∅ , where n is the number of

segments and R



(i, j) ∈ N

| I(i, j) = δ



, being I ( i, j ) the value

of the pixel located in ( i, j ) of the input image I and δ

is the kth

threshold value [25] . That is, the segmentation consists on group-

ing the pixels according to speciﬁc features of the object to recog-

nize; such as texture, shape, color, among others [173] . Segmen-

tation of images by color features has been addressed or stud-

ied recently. The algorithms for color image segmentation have

been developed because color features may provide relevant data

about the objects within the image. These algorithms have been

applied in different areas such as medicine [9,55,127,160,193] and

food analysis [47,108,111] , among others [7,14,33,74,137,169,174] .

Many of the techniques developed for image segmenta-

tion in gray scale have been extended for color images

[34,95,114,147,158,172,197] ; however, such techniques cannot be al-

ways successfully applied, because they are designed to process

∗

Corresponding author.

E-mail addresses: fgarcial@uaemex.mx (F. Garcia-Lamont),

jcervantesc@uaemex.mx (J. Cervantes), alchau@uaemex.mx (A. López),

lrodriguez@itorizaba.edu.mx (L. Rodriguez).

mainly the intensity of the colors without considering the chro-

maticity. Therefore, the algorithms for color image segmentation

must be developed taking into account the characteristics of color.

For color processing, it is important to select an adequate mathe-

matical representation of color, such that all the features of color

can be processed independently; basically, the most important fea-

tures of color are intensity and chromaticity [49] . There are differ-

ent color spaces to represent color; selecting a color space depends

on its characteristics, the way the color is contemplated to be pro-

cessed and the nature of the method employed for color process-

ing. For instance, the RGB color space is adequate for image dis-

playing, but not for color processing, because the intensity is not

decoupled from the chromaticity. Thus, in this paper we present

the color spaces more often used in related works, the character-

istics of these color spaces are described and we compare their

advantages and disadvantages. An important part of the segmen-

tation stage is the quantitative evaluation of the segmented image.

So far, there have not been deﬁned standard metrics for quanti-

tative evaluation of color image segmentation. Therefore, in this

survey we present a set of metrics for quantitative evaluation of

color image segmentation; these metrics are often employed to

evaluate quantitatively the segmentation of color images The rest

of this paper is divided as follows: in Section 2 the most common

color spaces employed for image segmentation are presented; in

addition their characteristics, advantages and disadvantages are de-

scribed. In Section 3 the segmentation techniques for color images

of previous works are introduced and reviewed. In Section 4 a set

https://doi.org/10.1016/j.neucom.2018.01.091

2 F. Garcia-Lamont et al. / Neurocomputing 292 (2018) 1–27

of metrics widely used for quantitative evaluation of color image

segmentation are presented. Section 5 shows some applications of

color image segmentation. Finally, Section 6 closes the paper with

conclusions.

2. Color spaces

The goal of a color space is to ease the speciﬁcation of colors

within a tridimensional coordinated system, or form a subspace of

the system where every color is represented by a unique point.

Most of the color spaces employed are oriented to hardware de-

vices, such as monitors, printers, or applications for color manip-

ulations, like creation of graphics for animation. The usual mod-

els oriented to hardware are RGB (red, green, blue) for monitors

and video cameras; CMY (cyan, magenta, yellow) for printers, and

YIQ (where Y is brightness and I and Q are chromatic components)

which is the standard for television [49] .

In the literature, the color spaces for color image processing are

the following: RGB, HSV (hue, saturation, value), HSI (hue, satura-

tion, intensity), L

∗

, L

∗

, YUV and YCbCr. Table 9 shows the

color spaces employed in different previous works to represent and

process colors.

The RGB space is adequate for color displaying, for instance, it

is widely employed for television systems and image acquisition;

although this is space is often employed for color recognition, the

RGB space is not suitable for segmentation or color processing, be-

cause of the high correlation between the components R, G and

B. There are other spaces that do not have this problem, but they

have their respective disadvantages. Next we present the features

of the RGB, HSV, HSI, L

∗

, L

∗

, YUV and YCbCr color spaces;

as well as, the equations to map colors between the RGB space and

the color spaces mentioned.

2.1. RGB space

In this space every color is represented with the spectral com-

ponents of red, green and blue. The origin of this model can be

found in television technology, and it can be considered as the

fundamental representation of color for computers, digital cameras

and scanners; but also, for image storage. Most of the software de-

veloped for image processing and graphics employ this model. In

the RGB model the combination of colors is based on the addition

of the individual components considering as base the black color.

The process can be considered as the combination of three rays

of color red, green and blue. The intensity of the different compo-

nents of color determines both the hue and the brightness of the

resulting color [49] . The shape of the RGB space is a cube, whose

coordinates correspond to the three basic colors: red (r), green (g)

and blue (b). The values of each component are in the range [0,

255] ⊂ , where every possible color corresponds to a point within

the cube; but, usually the range of values of each color compo-

nent are normalized to the range [0, 1]; hence, space color is rep-

resented as the unit cube shown in Fig. 1 .

The colors red, green and blue constitute the coordinate axis;

different colors are obtained by combining the values of the coor-

dinate axis. Table 2 shows the combination values of the axis for

usual colors.

The RGB space is a simple model, in several studies it is em-

ployed for color processing or when it is necessary to transform

colors to a different color space.

2.2. HSV space

In the HSV space, the color is represented with the components

hue (h), saturation (s) and value (v). Hue is the chromatic feature

that describes a pure color; for instance, yellow, orange, red, etc.

Fig. 1. RGB space represented as a unit cube.

Table 1

Color spaces employed for color image segmentation in different works.

Color space References

RGB [8,22,30,51,68,72,94,106,108,119,136,142,146,157,165,170]

HSV [30,67,69,100,105,118,137,150,189]

HSI [30,67,69,100,105,118,137,150,189]

∗

[22,30,66,81,94,139,185]

∗

[30,125,155,160,165,191,193]

YUV [23,26,27,80,181,186]

YCbCr [44,80,105,109,143,145]

Table 2

Usual colors and their corresponding red, green and blue values.

Color R G B

Black 0 0 0

Red 1 0 0

Yellow 1 1 0

Green 0 1 0

Cyan 0 1 1

Blue 0 0 1

Magenta 1 0 1

White 1 1 1

Gray

Table 3

Usual colors and their corresponding hue, saturation and value parameters.

Color H S V

Black Undeﬁned 0 0

Red 0 1 255

Yellow

1 255

Green

π 1 255

Cyan

π 1 255

Blue

π 1 255

Magenta

π 1 255

White Undeﬁned 0 255

Gray Undeﬁned 0 127

Table 4

Usual colors and their corresponding hue, saturation and intensity parameters.

Color H S V

Black Undeﬁned 0 0

Red 0 1 255

Yellow

1 255

Green

π 1 255

Cyan

π 1 255

Blue

π 1 255

Magenta

π 1 255

White Undeﬁned 0 255

Gray Undeﬁned 0 127

F. Garcia-Lamont et al. / Neurocomputing 292 (2018) 1–27 3

Table 5

Usual colors and their corresponding L

∗

, a

∗

and b

∗

parameters.

Color L

∗

Black 0 0 0

Red 100 128 128

Yellow 100 0 128

Green 100 −127 128

Cyan 100 −127 0

Blue 100 −127 −127

Magenta 100 128 0

White 100 0 0

Gray 50 0 0

Fig. 2. HSV color space.

Saturation is a measure of how the hue is diluted in white light;

value is the intensity or brightness of the color. The HSV space has

two important features [49,69] :

1. The intensity component, value, is decoupled from the hue

data;

2. The hue and saturation components emulate the human per-

ception of color. With these features the HSV space becomes

a useful tool to develop image processing algorithms based on

some properties of the human color perception.

The hue is in the range [0, 2 π ] ⊂ ; saturation is in the real

range [0, 1], while value is often in the range [0, 255] ⊂ . The HSV

space is cone shaped, as shown in Fig. 2 . Geometrically, the radius

and the height of the cone represent the saturation and value com-

ponents, respectively.

Note that for colors black, white and gray, the hue parameter

is undeﬁned, due to these colors are considered as singularities

within this color space; because they do not have a speciﬁc chro-

maticity.

2.3. HSI space

The HSI space, represented with the components hue (h), sat-

uration (s) and intensity (i), is very similar to the HSV space. De-

spite the similarity, computing their components, hue, saturation

and the intensity, is different with respect to how same compo-

nents are obtained in the HSV space. The ranges of the components

are the same than the ranges of the components of the HSV space.

That is, the hue is in the range [0, 2 π] ⊂ ; saturation is in the

real range [0, 1], while intensity is in the range [0, 255] ⊂ . The

Fig. 3. HSI color space.

HSI space is double cone shaped, as shown in Fig. 3 . Geometrically,

the radius and the height of the cone represent the saturation and

intensity components, respectively [49] .

For both spaces HSV and HSI, the same hue and saturation pa-

rameters are employed to represent colors, except for brightness,

where in each space the parameters are different.

2.4. CIE XYZ space

The standard model XYZ, developed by the Comisin Interna-

tionale dclairage, is the base of most of the calibrated color models

employed nowadays. The calibrated color models are employed to

reproduce the colors independently of the display devices. Several

problems happen because there is a strong dependence between

the device and the reproduction of images. All the color spaces

described before are related to the physic measures of the output

devices employed to display images; for instance, the conﬁgurable

parameters of a laser printer. The color model is developed by sev-

eral measures performed under strict conditions. The model con-

sists of three basic colors X, Y and Z; they are selected such that,

through positive components all the colors and combinations can

be described. This space is perceived as no linear by humans. That

is, in some parts of the space, huge color changes are produced

before little position variations; while in other parts of the space

happens the opposite, little color changes are experimented for

large position changes [49] . Thus, variations of the CIE model have

been developed for different kinds of applications, or to repre-

sent the colors such that it mimics the human perception of color.

Examples of CIE variations are the spaces YUV, YCbCr, L

∗

and

∗

; in this study we address the spaces L

∗

, L

∗

, YUV

and YCbCr because they are often employed for color image pro-

cessing.

2.4.1. CIE L

∗

space

This color space is developed considering linearizing the tonal-

ity changes, where the colors are deﬁned by three variables: L

∗

the intensity, a

∗

and b

∗

are the tonality components [49,66] . The

value of a

∗

deﬁnes the distance through red-green axis, while the

value of b

∗

deﬁnes the distance through the blue-yellow axis. Usu-

ally a

∗

and b

∗

are in the ranges [ −127 , 128] ⊂ and L

∗

in the range

[0, 100] ⊂ .

The shape of this space is similar to the RGB space, but the

location of colors is different, see Fig. 4 .

Note that in this space, similarly to the HSV and HSI spaces, the

intensity is decoupled from the chromaticity, mimicking the way

4 F. Garcia-Lamont et al. / Neurocomputing 292 (2018) 1–27

Fig. 4. L

∗

color space.

Table 6

Usual colors and their corresponding L

∗

, u

∗

and v

∗

parameters.

Color L

∗

Black 0 0 0

Red 53 175 37

Yellow 94 7 106

Green 84 −83 107

Cyan 88 −70 −15

Blue 29 −9 −130

Magenta 57 80 −108

White 100 0 0

Gray 50 0 0

Table 7

Usual colors and their corresponding Y, U and V parameters.

Color Y U V

Black 0 0 0

Red 0.299 −0.147 0.615

Yellow 0.886 −0.436 0.1

Green 0.587 −0.289 −0.515

Cyan 0.701 0.147 −0.615

Blue 0.114 0.436 −0.1

Magenta 0.413 0.289 0.515

White 1 0 0

Gray 0.5 0 0

the humans perceive the colors

[69] . But also, because of the tonal-

ity changes are linear, the chromaticity differences can be com-

puted using the Euclidean distance.

2.4.2. CIE L

∗

space

This space is similar to the CIE L

∗

space, where the colors

are represented by the intensity component L

∗

and the chromatic

components u

∗

and v

∗

[49] . Usually u

∗

and v

∗

are in the ranges

[ −175 , 175] ⊂ and L

∗

in the range [0, 100] ⊂ . Table 6 shows the

parameters for luminance, and chromaticity for usual colors.

2.4.3. YUV and YCbCr color spaces

The YUV and YCbCr color spaces are employed to standardize

the images for television. The YUV space is the base for color cod-

ing for NTSC system; while the YCbCr is the standard for digi-

tal television. In these models the components that deﬁne them

feature three planes: the luminance (Y) and the other two called

chrominance components, (UV and CbCr for spaces YUV and YCbCr,

respectively) [49] . Table 7 shows the luminance and chrominance

parameters for usual colors for the YUV space.

Due to humans are not able to distinguish sharpness with high

precision in colors, and also humans are more sensitive to bright-

ness, the bandwidth can be reduced considerably for the deﬁnition

of the color components. This feature is employed by the color

Table 8

Usual colors and their corresponding Y, Cb and Cr parameters.

Color Y Cb Cr

Black 0 0 0

Red 0.299 −0.169 0.5

Yellow 0.886 −0.5 0.081

Green 0.587 −0.331 −0.419

Cyan 0.701 −0.831 −0.5

Blue 0.114 −0.5 −0.081

Magenta 0.413 −0.669 0.419

White 1 −1 0

Gray 0.5 −0.5 0

compression algorithms; for instance, as a part of the compres-

sion algorithm for the JPEG format, the RGB colors are mapped

to the YCbCr space. However, the YCbCr space is widely employed

for image processing, mainly for compression applications, the YUV

model is less used. Table 8 shows the luminance and chrominance

paramaters for usual colors for the YCbCr space.

2.5. Transformations between color spaces

As mentioned before, the hardware devices employed for image

acquisition and display, employ the RGB space to represent colors.

Therefore, it is necessary to map the colors to the color spaces pre-

sented above, in order to process the colors under the features of

such color spaces; and then, to map the resulting colors to the RGB

space so as to display the result of processing. In this section, we

present the equations to map the RGB colors to the color spaces

mentioned before, as well as the inverse mapping [25,49] .

2.5.1. Mapping between RGB and HSV spaces

Mapping a RGB color to the HSV space involves the following

operations. Let ϕ = [ r, g, b] be the RGB color vector, and φ = [ h, s, v ]

the resulting vector by mapping ϕ to the HSV space. The hue, sat-

uration and value are computed with:

θ =

⎧

⎪

⎨

⎪

⎩

und ef ined , r = g = b

cos

−1



(

r − g

)

(

r − b

)



(

r − g

)

(

r − b

) (

g − b

)



, ot herwise

(1)

h =



, b ≤ g

2 π − θ, b > g

(2)

s =



0 , max (r, g, b) = 0

1 −

min

(r, g, b)

max (r, g, b)

, ot her wise

(3)

v = max (r, g, b) (4)

The inverse operation, in other words, mapping HSV color vec-

tor φ = [ h, s, v ] to the RGB space involves the following operations.

If h = und ef ined then r = v , g = v and b = v ; otherwise:

r =

⎧

⎪

⎨

⎪

⎩

q, k = 1

p, 2 ≤ k ≤ 3

t, k = 4

v , ot her wise

(5)

g =

⎧

⎪

⎨

⎪

⎩

t, k = 0

v , 1 ≤ k ≤ 2

q, k = 3

p, ot herwise

(6)

b =

⎧

⎪

⎨

⎪

⎩

p, 0 ≤ k ≤ 1

t, k = 2

v , 3 ≤ k ≤ 4

q, ot herwise

(7)

F. Garcia-Lamont et al. / Neurocomputing 292 (2018) 1–27 5

where

k =





(8)

f =

h − k (9)

p = v × (1 − s ) (10)

q = v ×

(

1 − (s × f )

)

(11)

t = v ×

(

1 − (s × (1 − f ))

)

(12)

2.5.2. Mapping between RGB and HSI spaces

Mapping a RGB color to the HSI space involves the following

operations. Let ϕ = [ r, g, b] be the RGB color vector, and ψ = [ h, s, i ]

the resulting vector by mapping ϕ to the HSI space. The hue is

computed with Eqs. (1) and (2), while saturation and intensity are

computed with:

s = 1 −

min

(r, g, b)

(13)

i =

r + g + b

(14)

The inverse operation, in other words, mapping HSI color vector

ψ = [ h, s, i ] to the RGB space involves the following operations. If

h = und ef ined then r = i, g = i and b = i else: If 0 ≤ h <

2 π

then

b = i (1 − s ) (15)

r = i

⎡

⎣

1 +

s cos h

cos (

− h )

⎤

⎦

(16)

g = 3 i (1 − r − b) (17)

2 π

≤ h <

4 π

then

r = i (1 − s ) (18)

g = i

⎡

⎢

⎣

1 +

s cos



h −



cos (π − h )

⎤

⎥

⎦

(19)

b = 3 i

(

1 − r − g

)

(20)

π ≤ h < 2 π then

g = i (1 − s ) (21)

b = i

⎡

⎢

⎣

1 +

s cos



h −



cos



π − h



⎤

⎥

⎦

(22)

r = 3 i

(

1 − g − b

)

(23)

2.5.3. Mapping between RGB and L

∗

spaces

Mapping a RGB color to the L

∗

is performed with the fol-

lowing equations. Let ϕ = [ r, g, b] be a color represented in the RGB

space:

∗

= 116 f



re f



− 16 (24)

∗

= 500





re f



− f



re f



(25)

∗

= 200





re f



− f



re f



(26)







0 . 4124 0 . 3575 0 . 1804

0 . 2126 0 . 7151 0 . 0721

0 . 0193 0 . 1191 0 . 9502

 



(27)

where δ =

and

f (t) =



√

t , t > δ

3 δ

, t ≤ δ

(28)

The inverse operation is performed with the following opera-

tions:

X = X

re f



∗

+ 16

116

∗

500



(29)

Y = Y

re f



∗

+ 16

116



(30)

Z = Z

re f



∗

+ 16

116

−

∗

200



(31)

where

h (t) =



, t > δ

3 δ



t −



, t ≤ δ

(32)







0 . 4124 0 . 3575 0 . 1804

0 . 2126 0 . 7151 0 . 0721

0 . 0193 0 . 1191 0 . 9502



−1





(33)

The values X

ref

, Y

ref

and Z

ref

are obtained by substituting the

tristimulus values for the reference white.

2.5.4. Mapping between RGB and L

∗

v spaces

Mapping a RGB color to the L

∗

is obtained with the follow-

ing equations. Let Ï,g,b] be a color represented in the RGB space:

∗

⎧

⎪

⎨

⎪

⎩







re f





re f



≤ δ

116



re f



− 16 ,



re f



> δ

(34)

∗

= 13 L

∗





− u



re f



(35)

∗

= 13 L

∗





− v



re f



(36)



4 X

X + 15 Y + 3 Z

(37)



9 Y

X + 15 Y + 3 Z

(38)

剩余26页未读，继续阅读

qq_28368825

2019-06-20

怎么下载只有864B

qq_38308100

粉丝: 0
资源: 3

会员权益专享

一篇图像分割的外文论文，可供大家学习

评论1

会员权益专享

最新资源

一篇图像分割的外文论文，可供大家学习

评论1

我的毕设是基于无线供能的传感物联网系统设计与实现，请帮我找20篇相关的外文论文以及网址

Spring相关的外文文献和翻译(毕设论文必备).doc

帮我找五篇决策树的外文论文

给我一篇计算机科学相关的外文文献

帮我找一篇关于STM32智慧农业系统设计与实现论文的外文翻译

帮我找一篇关于STM32智慧农业系统设计与实现论文的外文翻译的链接

帮我找一篇关于基于单片机的电车电瓶防盗报警器设计与实现的论文外文翻译的链接

物流 外文翻译 毕业论文 sql

计算机毕业设计的外文翻译（学长去年的论文）

SSM外文文献和翻译(毕设论文精品).doc

帮我找一篇关于基于单片机的电车电瓶防盗报警器设计与实现的外文翻译的链接

计算机专业外文文献翻译

与语义协同相关的外文文献

ssm框架外文文献翻译5000字

请提供几篇关键词为工程机械、gazebo、仿真的外文文献

spring框架外文文献.docx

帮我检索20篇有关语义协同的外文文献

基于STM32的健康监测仪设计的外文文献

C 代码 进行分子动力学 （MD） 模拟 并行使用 OpenMP.rar

会员权益专享

最新资源

物流外文翻译毕业论文 sql

C 代码进行分子动力学（MD）模拟并行使用 OpenMP.rar