Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more ➡
Download
Standard view
Full view
of .
Add note
Save to My Library
Sync to mobile
Look up keyword
Like this
1Activity
×
0 of .
Results for:
No results containing your search query
P. 1
Image Retrieval with Image Tile Energy Averaging using Assorted Color Spaces

Image Retrieval with Image Tile Energy Averaging using Assorted Color Spaces

Ratings: (0)|Views: 277|Likes:
Published by ijcsis
Here the feature vector for image retrieval is composed of average energy of each tile of image for diverse number of image tiles (like 1, 4, 9, 16, 25, 36 and 49) considered with the help of various color spaces. The paper presents exhaustive performance comparison of 70 variants of proposed image retrieval technique using ten sundry color spaces and seven image tiling methods is done with the help of generic image database having 1000 images spread across 11 categories. For each proposed CBIR technique 55 queries (randomly selected 5 per category) are fired on the generic image database. To compare the performance of image retrieval techniques, average precision and recall are computed and plotted against number of retrieved images. The results have shown that RGB and HSI color spaces give the best performance for average energy based image retrieval across all tiles. Also it has been seen in all luminance-chromaticity based color spaces ( Kekre’s LUV, YCbCr, YUV, YIQ and Kekre’s YCgCb) that as the number tiles increased the overall performance also increases.
Here the feature vector for image retrieval is composed of average energy of each tile of image for diverse number of image tiles (like 1, 4, 9, 16, 25, 36 and 49) considered with the help of various color spaces. The paper presents exhaustive performance comparison of 70 variants of proposed image retrieval technique using ten sundry color spaces and seven image tiling methods is done with the help of generic image database having 1000 images spread across 11 categories. For each proposed CBIR technique 55 queries (randomly selected 5 per category) are fired on the generic image database. To compare the performance of image retrieval techniques, average precision and recall are computed and plotted against number of retrieved images. The results have shown that RGB and HSI color spaces give the best performance for average energy based image retrieval across all tiles. Also it has been seen in all luminance-chromaticity based color spaces ( Kekre’s LUV, YCbCr, YUV, YIQ and Kekre’s YCgCb) that as the number tiles increased the overall performance also increases.

More info:

Published by: ijcsis on Apr 10, 2011
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See More
See less

04/10/2011

pdf

text

original

 
(IJCSIS) International Journal of Computer Science and Information Security,Vol.
9
, No.
3
, 201
1
 
Image Retrieval with Image Tile Energy Averagingusing Assorted Color Spaces
Dr. H.B.Kekre
1
, Sudeep D. Thepade
2
, Varun Lodha, Pooja Luthra, Ajoy Joseph, Chitrangada Nemani
3
 
1
Senior Professor,
2
Ph.D.Research Scholar & Associate Professor,
3
B.Tech (IT) Student
 
Information Technology Department, MPSTME,
SVKM‟s
NMIMS (Deemed-to-be University), Mumbai, India
1
hbkekre@yahoo.com,
2
sudeepthepade@gmail.com,
3
ajoy.jose90@gmail.com,
 
chitrangada.nmims@gmail.com,
 
varunlodha_4@hotmail.com,
 
poojapoohluthra@gmail.com
Abstract
— 
 
Here the feature vector for image retrieval is composedof average energy of each tile of image for diverse number of imagetiles (like 1, 4, 9, 16, 25, 36 and 49) considered with the help of various color spaces. The paper presents exhaustive performancecomparison of 70 variants of proposed image retrieval techniqueusing ten sundry color spaces and seven image tiling methods isdone with the help of generic image database having 1000 imagesspread across 11 categories. For each proposed CBIR technique 55queries (randomly selected 5 per category) are fired on the genericimage database. To compare the performance of image retrievaltechniques, average precision and recall are computed and plottedagainst number of retrieved images. The results have shown thatRGB and HSI color spaces give the best performance for averageenergy based image retrieval across all tiles. Also it has been seen inall luminance-
chromaticity based color spaces ( Kekre’s LUV,YCbCr, YUV, YIQ and Kekre’s YCgCb) that as the number tiles
increased the overall performance also increases.Keywords: CBIR, Average Energy, Color Spaces, Image Tiling.
I.
 
I
NTRODUCTION
 The large numbers of images which are being generated from avariety of sources (digital camera, digital video, scanner, theinternet etc.) have posed technical challenges for computersystems to store/transmit and index/manage image dataeffectively to make such collections easily accessible. Imagecompression deals with the challenge of storage andtransmission, where significant advancements have been made[1,4,5]. The challenge to image indexing is studied in the contextof image database [2,6,7,10,11], which has become one of thepromising and important research area for researchers from awide range of disciplines like computer vision, image processingand database areas. The thirst for better and faster imageretrieval techniques is increasing day by day. Problems withtraditional methods of image indexing have led to the rise intechniques for retrieving images on the basis of automaticallyderived features such as color, texture and shape- a technologynow referred as Content-Based Image Retrieval (CBIR). Someof the important applications for CBIR technology could beidentified as art galleries [12,14], museums, archaeology [3],architecture design [8,13], geographic information systems [5],weather forecast [5,22], medical imaging [5,18], trademark databases [21,23], criminal investigations [24,25], image searchon the Internet [9,19,20].A Content Based Image Retrieval (CBIR) is an interfacebetween a high level system (the human brain) and a low levelsystem (a computer). The human brain is capable of performingcomplex visual perception, but is limited in speed while acomputer is capable of limited visual capabilities at much higherspeeds. In a CBIR, features are used to represent the imagecontent. The features are extracted automatically and there is nomanual intervention, thus eliminating the dependency on humansin the feature extraction stage. These automated approaches toobject recognition are computationally expensive, difficult andtend to be domain specific. The typical CBIR system performstwo major tasks [16,17]. The first one is feature extraction (FE),where a set of features, called feature vector, is generated toaccurately represent the content of each image in the database.The second task is similarity measurement (SM), where adistance between the query image and each image in thedatabase using their feature vectors is used to retrieve the top
“closest” images [16,17,26]. For feature extraction in CBIR there
are mainly two approaches [5] feature extraction in spatialdomain and feature extraction in transform domain. The featureextraction in spatial domain includes the CBIR techniques basedon histograms [5], BTC [1,2,16], VQ [21,25,26]. The transformdomain methods are widely used in image compression, as theygive high energy compaction in transformed image [17,24]. So itis obvious to use images in transformed domain for featureextraction in CBIR [23]. But taking transform of image is timeconsuming and also needs all images of database to be of samesize to get similar feature vectors. This limitation is overcomehere in proposed CBIR methods using average energy conceptwith help of image tiling.II.
 
CONSIDERED
 
COLOR
 
SPACES
 
Including RGB color space, in all ten assorted color spaces areconsidered here.
A.
 
Kekre’s LUV Color Space
 
Kekre‟s LUV color Space is special case of Kekre Transform.
Where L gives luminance and U and V gives chromaticity valuesof color image. Positive value of U indicates prominence of redcomponent in color image and negative value of V indicatesprominence of green component. This needs the conversion of 
280 http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol.
9
, No.
3
, 201
1
 
RGB to LUV components. The RGB to LUV conversion matrixgiven in equation 1 gives the L, U, V components of color imagefor respective R, G, B components.
BGRL
.110112111 
 
(1)
 
The LUV to RGB conversion matrix given in equation 2 givesthe R, G, B components of color image for respective L, U, Vcomponents.
263.111111021 
LBGR
 
(2)
B.
 
YCbCr Color Space
In YCbCr color Space, Y gives luminance and Cb and Cr giveschromaticity values of color image. To get YCbCr componentswe need the conversion of RGB to YCbCr components. TheRGB to YCbCr conversion matrix given in equation 3 gives theY, Cb, Cr components of color image for respective R, G, and Bcomponents.
BGRCr Cb
.0816.04184.00.5000 5000.03312.01688.0 1145.05866.02989.0  
 
(3)
 
The YCbCr to RGB conversion matrix given in equation 4 givesthe R, G, B components of color image for respective Y, Cb, andCr components.
Cr CbBGR
.0010.07718.11 7140.03441.01 4020.10010.01  
 
(4)
 
C.
 
YUV Color Space
The YUV model defines a color space in terms of one luminance(brightness) and two chrominance (color) components. The YUVcolor model is used in the PAL, NTSC, and SECAM compositecolor video standards. Previous black-and-white systems used
 
only luminance (Y) information and color information (U and V)was added so that a black-and-white
 
receiver would still be
 
ableto display a color picture as a normal black and white pictures.YUV models human perception of color in a different way thanthe standard RGB model used in computer graphics hardware.The human eye has fairly little color sensitivity: the accuracy of the brightness information of the luminance channel has far moreimpact on the image discerned than that of the other two. TheRGB to YUV conversion matrix given in equation 5 gives the Y,U, V components of color image for respective R, G, Bcomponents.
BGR
.10001.051498.00.615 436.022472.014713.0 144.0587.0299.0  
 
(5)
 
The YUV to RGB conversion matrix given in equation (6) givesthe R, G, B components of color image for respective Y, U, Vcomponents.
BGR
.000015.09729.10.97086 5876.022472.00836.1 1398.150901.07492.0  
 
(6)
 
D.
 
YIQ Color Space
The YIQ color space is derived from YUV color space and isoptionally used by the NTSC composite color video standard.
The ‟I‟ stands for in phase and „Q‟ for quadrature, which is the
modulation method used to transmit the color information.
BGRQ
.31135.0522591.00.211456 321263.0274453.0595716.0 144.0587.0299.0  
 
(7)
 
The inter-conversion equations for YIQ to RGB color space aregiven as per the equations (7) and (8).
QBGR
.7046.1107.11 6474.02721.01 6210.09563.01  
 
(8)
 
E.
 
Kekre’s YCgCb Color Space
Inter-
conversion equations for RGB to Kekre‟s YCgCb color 
space can be given as below in equations 9 and 10.
BGRCbCg
.101011111 
 (9)
223.101011111 
CbCgBGR
 (10)
F.
 
XYZ Color Space
Conversion equations for RGB to XYZ color space and XYZ toRGB can be given as given in equations 11 and 12 below.
BGR
.950227.0119193.00.019334 072169.071160.0212671.0 180423.0357580.0412453.0  
 
(11)
BGR
.057311.1204043.00.055648 041556.0875992.1969256.0 498535.0537150.1240479.3  
 
(12)
G.
 
rgb Color Space (Normalized RGB)
In order to eliminate the influence of illumination intensity, colorinformation (R, G and B) can be normalized to get rgb colorspace where,
            
 
(13)
281 http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol.
9
, No.
3
, 201
1
 
H.
 
HSV Color Space
The HSV stands for the Hue, Saturation and Value based on theartists (Tint, Shade, and Tone). The Value represents intensityof a color, which is decoupled from the color information in therepresented image. The Hue and Saturation components areintimately related to the way human eye perceives color resultingin image processing algorithms with physiological basis.Conversion formula from RGB to HSV is as follows.
  

    
  
 
(14)
      
 
(15)
   
 
(16)
Conversion from HSV space to RGB space is more complex.And, given to the nature of the hue information, we will have adifferent formula for each sector of the color triangle.
Red-Green Sector:
for
    
 
(17)
         
 
Green-Blue Sector:
for
    
 
(18)
           
 
Blue-Red Sector:
for
    
 
(19)
         
 
I.
 
HSI Color Space
To convert RGB to HSI [29,30], first we convert RGB to
„normalized rgb‟ using equations given in section 2.2.G. Each
normalized H, S and I are then obtained using followingequations
.
gforbhbggg bg h
,,0, 212211cos
 
(20)
gbfor h bggg bg h
,,2,, 212211cos2
 
(21)
1,0,,,min.31
sbgs
 
(22)
1,0,)255.3/()(
iBGRi
 
(23)
For convenience h,s and i values are converted in the ranges of [0,360],[0,100] and [0,255] respectively using followingequation 24.
255,100,180
isSh
 
(24)
III.
 
IMAGE
 
TILING
 
[24]Tiling of an image is basically dividing an image into differentequal sized, non overlapping quadrants for feature extraction.Here seven assorted image tiling techniques are applied onimages for feature extraction per colour space in proposed CBIRmethods.IV.
 
PROPOSED
 
CBIR
 
TECHNIQUESEnergy averaging could be defined as saverage of squared valuesof pixels of the respective image tile. Here average energytechnique is applied along with tiling for feature vectorgeneration. Feature vector size for respective image tilingmethod is shown in table 1.
 
Image Tiling Method1-Tile(1x1)4-Tile(2x2)9-Tile(3x3)16-Tile(4x4)25-Tile(5x5)36-Tile(6x6)49-Tile(7x7)Number of Tiles 1 4 9 16 25 36 49Feature vector Size 3 12 27 48 75 108 147
Table 1 Feature vector size for respective image tiling method considered
282 http://sites.google.com/site/ijcsis/ISSN 1947-5500

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->