Row Mean and Column Mean Based Cbir System for Gray Scale Image

Row Mean and Column Mean Based Cbir System for Gray Scale Image

Madhavi kshatri 1, Yogesh Ratore 2,

1 Department of CSE., RITEE College, Raipur(Chhattisgarh.),INDIA

2 Department of CSE ., RITEE,Raipur(Chhattisgarh),INDIA

Abstract

CBIR is a technique which is used to index and search the relevant images from the large image database as per the user requirement. For this the CBIR system relies on the visual contents of the image .Since the intensity distribution of pixels in similar or relevant images are approximately similar therefore this characteristics can be utilized to design an efficient CBIR system.This paper present a unique CBIR system which utilizes the average intensity of the pixels present in the grayscale image.

1. Introduction

   The increasing amount of digital content In the form of audio video and images in the web and other digital media has lead us to device some application to search effectively for relevant information in existing and increasing database of image ,audio and video content.CBIR is one such kind of system in which images can be indexed effectively by summarizing their visual feature.A visual feature is a characteristic property of an image that can be used to represent the whole image approximately with reduced dimension so that the searching and browsing images in image data base can become easy with less storage space and computational power.Such features can be globally for whole image or locally for some part of the image or object. Texture, shape and color are commonly used feature in any image retrieval system.

   Since every image has unique color ,texture and shape information therefore images can be retirives using these information.In the past image retrieval; system based on color [1] [2], shape[2][3][4],

   texture[5] has been proposed. J. Berens, G. D. Finlayson and G. Qiu [6] presented color histogram based retrieval system in compressed transform domain.

   Image retrieval system based on color histogram[7][8] and by computing color average[9][10] were also presentedimage retrieval system based on discrete cosine transform [11] [12] and wavelet transform were also proposed in the past.

   In some images shape or shape of a region are also very important features.image segmentation and edge detection can be used to describe the shape related information in an image and hence can be used [13]for image retrieval purposes.This paper present a image retrieval system which is based on Row sum , column sum. Since the intensity of the pixel in an image carry very vital information about the content of the image, therefore these information can be used to design a very efficient image retrieval system .

2. Proposed CBIR system

   1. Image database creation

      proposed system architecture is given in the figure- 1.in order to reduce the computational complexity for indexing and retrieval process, gray level image of size 256*256 has been taken.As shown in figure the first step for making data base is to take each gray scale image, resize it to 256*256 and then extract the rows, column and diagonal of image in separate matrix and then computing the row mean,column mean, diagonal mean and store it in a data base which has three field for each image i.e. DM,RM,CM for diagonal mean,row mean and column mean respectively.gray scale images of different field taken

      Input Query Image
      	Resize Image 256*256

      personally and from internet is used to make a database for this proposed system.

      M=No. of rows N=No. of Column

      CM (j) =

      ..(1)

      RM (i) =

      Extract Column,Row

      Compute Column Mean

      Compute Row Mean

      ,Diagonal

      CM

      RM

      CM

      RM

      CM

      RM

      Img n

      .

      Img2

      Img1

      Output Image

      CM

      RM

      Compute Column Mean

      Compute Row Mean

      Resize Image 256*256

      Extract Column,Row,Diagonal

      Similarity Measurement

      Img1		Img2		.	Img n
      RM	CM	RM	CM		RM	CM
      Image Feature Data Base

      Image Feature Data Base

      Figure.1 Block diagram of database Creation

   2. Image Retrieval System-

      Image retrieval system for proposed CBIR is shown in figure 2. First of all the query image is inputed to the retrieval system which resize it to 256*256 and then after extracting the rows,column, diagonal of the image,it compute the Diagonal Mean, Row mean,

      Figure.2 Block Diagram of Proposed CBIR

3. Similarity Measurement-

   For similarity measurement ,Euclidean Distance is computed between the query image and Database image using Formula(3)

   Column Mean and then compare this DM,RM and CM of Query image to the DM,RM and CM of the Images stored already in the image database and

   compute the similarity measurement, image or images which has similarity coefficient less than predefined threshold T are the similar images.

   Input Query Image

   FQI = feature vector of Query Image FDI = feature vector of Database Image n=No. of features

4. Threshold Selection-

   For this system to work properly,a suitable threshold T selection is very important.By considering Maximum distance between query image and database image, as a reference point,an extensive experimentation is carried out using four different threshold,

   First threshold has been chosen as 15% of the maximum ED, second threshold as 25% of maximum ED and third threshold as 35% of maximum ED.

   Out of which the threshold of 25% of maximum ED was found giving better result and therefore selected as the defined threshold T for this algorithm.

5. Experimental Result

   The proposed image retrieval algorithm is implemented using MATLAB 7.0 on a system with core 2 duo processor and 2GB of RAM.To verify the performance of proposed method, a database of 1000 images taken from coral collection has been used.Data base images has been divided into 8 different categories as shown in table1.

   Table 1. Image Category

   Precision

   1

   0.8

   0.6

   0.4

   0.2

   0

   0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

   Recall

   Figure.3 Precision Vs Recall Curve for proposed method

   0.8

   Average Precision

   0.7

   0.6

   0.5

   0.4

   Group ID	Image category
   1	Bus
   2	Dinosaur
   3	Elephant
   4	Tribal /td>
   5	Beaches
   6	Flower
   7	Horse
   8	Mountain

   0.3

   0.2

   0.1

   0

   0 1 2 3 4 5 6 7 8 9 10 11

   Image category

   Figure.4 Average Precision Curve for proposed method

   All the images are converted to gray scale and stored in JPEG format with size 256X256.Five different images of each category are used for query images. Euclidean Distance has been computed as given by equation 3.All the images for which the distances of query image from the database images are less than predefined threshold T has been considered as a retrieved image.To check the efficiency of our proposed work,two statistical parameter viz. precision and recall as given by equation 4 and 5 has been computed.Precision versus recall curve has been drawn and shown in figure 3 while

   Average precision curve is shown in figure 4.

   Figure 5 GUI of proposed method

   Figure 6 Query image (upper part) and first 10 retrieved images(lower part).

6. Conclusion

The performance of image retrieval system can be judged by the efficiency and speed at which it retrived the relevant images from the large database.Since the proposed system uses column sum and row sum for image retrieval which require less computation power and hence it is faster moreover from experimental result it is clear that the proposed system is efficient as well.

References

1. Hossein Nezamabadi-pour and Saeid Saryazdi,

   Object- Based Image Indexing and retrieval in DCT Domain using Clustering Techniques, Vol. 3 JANUARY 2005 ISSN 1307-

   6884.

2. M. J. Swain and D. H. Ballard, Color

   indexing, International Journal of Computer Vision, 1991, vol.7, no.1, pp.11-32.

3. A. K. Jain and A. Vailaya, Image retrieval using color and shape, Pattern Recognition, 996, vol.29, no.8, pp.1233-1244.

4. F. Mokhtarian and S. Abbasi, Shape

   similarity retrieval under affine transforms, Pattern Recognition, 2002, vol. 35, pp. 31-41.

5. B. S. Manjunath and W. Y. Ma, Texture feature for browsing and retrieval of image data, IEEE PAMI, 1996, no. 18, vol. 8, pp. 837-842.

6. J. Berens, G. D. Finlayson and G. Qiu, Image indexing using compressed color Histograms, IEEE Proc.-Vision Image Signal Process. Vol. 147, No. 4, August 2000.

[7]. Hirata K. and Kato T. Query by visual example

content-based image retrieval, In Proc. Of Third International Conference on Extending Database Technology, EDBT92, 1992, pp 56-

71.

[8]. John Eakins, Margaret Graham, Content

Based Image Retrieval, Chatpter 5.6, pg 36- 40, University of Northrumbia at New Castle,

October

[9]. Dr.H.B.Kekre, Sudeep D. Thepade, Archana Athawale, Anant Shah, Prathmesh Verlekar, Suraj Shirke,Energy Compaction and Image Splitting for Image Retrieval using Kekre Transform over Row and Column Feature Vectors, International Journal of Computer Science and Network Security (IJCSNS),Volume:10, Number 1, January 2010, (ISSN: 1738-7906) Available at www.IJCSNS.org.

[10]. Dr.H.B.Kekre, Sudeep D. Thepade, Varun K. Banura, Amelioration of Colour Averaging Based Image Retrieval Techniques using Even

and Odd parts of Images, International Journal of Engineering Science and Technology (IJEST), Vol. 2, Issue 9, Sept. 2010. pp. (ISSN:

0975-

1. Jose A. Lay and Ling Guan, Image Retrieval Based on Energy

   Histograms of the Low Frequency DCT Coefficients, IEEE 0-7803-5041-3/99, 1999.

2. Stepan Obdrzalek and Jiri Matas, Image Retrieval Using Local Compact DCT-based Representation. DAGM03, 25th Pattern Recognition Symposium, September 10-12, 2003.

[13]. Sagarmay Deb, Yanchun Zhang, An Overview of Content Based Image Retrieval Techniques, Technical Report, University of Southern Queensland.

Extract Column,Row

,Diagonal

Resize Image 256*256

Resize Image 256*256

Extract Column,Row

,Diagonal

Compute Diagonal Mean

Compute Row Mean

Compute Column Mean

Img1 Img2 . Img n

DM RM CM DM RM CM DM RM CM

CM

RM

DM

Compute Column Mean

Compute Row Mean

Compute Diagonal Mean

Image Feature Data Base

Similarity Measurement

Output Image

Img1 Img2 . Img n

DM RM CM DM RM CM DM RM CM

Image Feature Data Base

EXPERIMENTAL RESULT-

3. Results and Discussion

Input Query Image

References

1. Hossein Nezamabadi-pour and Saeid Saryazdi, Object-Based Image Indexing and Retrieval

   in DCT Domain using Clustering Techniques, Vol. 3 JANUARY 2005 ISSN 1307-6884.

2. M. J. Swain and D. H. Ballard, Color

   indexing, International Journal of Computer Vision, 1991, vol.7, no.1, pp.11-32.

3. A. K. Jain and A. Vailaya, Image retrieval using color and shape, Pattern Recognition, 996, vol.29, no.8, pp.1233-1244.

4. F. Mokhtarian and S. Abbasi, Shape similarity retrieval under affine transforms, Pattern Recognition, 2002, vol. 35, pp. 31-41.

5. B. S. Manjunath and W. Y. Ma, Texture feature for browsing and retrieval of image data, IEEE PAMI, 1996, no. 18, vol. 8, pp. 837-842.

6. J. Berens, G. D. Finlayson and G. Qiu, Image indexing using compressed color

Histograms, IEEE Proc.-Vision Image Signal Process. Vol. 147, No. 4, August 2000.

[7]. Hirata K. and Kato T. Query by visual example

content-based image retrieval, In Proc. Of Third International Conference on Extending Database Technology, EDBT92, 1992, pp 56-

71.

[8]. John Eakins, Margaret Graham, Content

Based Image Retrieval, Chatpter 5.6, pg 36- 40, University of Northrumbia at New Castle,

October

[9]. Dr.H.B.Kekre, Sudeep D. Thepade, Archana Athawale, Anant Shah, Prathmesh Verlekar, Suraj Shirke,Energy Compaction and Image Splitting for Image Retrieval using Kekre Transform over Row and Column Feature Vectors, International Journal of Computer Science and Network Security (IJCSNS),Volume:10, Number 1, January 2010, (ISSN: 1738-7906) Available at www.IJCSNS.org.

[10]. Dr.H.B.Kekre, Sudeep D. Thepade, Varun K. Banura, Amelioration of Colour Averaging Based Image Retrieval Techniques using Even and Odd parts of Images, International Journal of Engineering Science and Technology (IJEST), Vol. 2, Issue 9, Sept. 2010. pp. (ISSN:

0975-

1. Jose A. Lay and Ling Guan, Image

   Retrieval Based on Energy Histograms of the Low Frequency DCT Coefficients, IEEE 0- 7803-5041-3/99, 1999.

2. Stepan Obdrzalek and Jiri Matas, Image Retrieval Using Local Compact DCT-based

Representation. DAGM03, 25th Pattern Recognition Symposium, September 10-12, 2003.

[13]. Sagarmay Deb, Yanchun Zhang, An Overview of Content Based Image Retrieval Techniques, Technical Report, University of Southern Queensland.

The image database was represented using a set of image

attribute, such as color [3] [4], shape [4] [5] [6], texture [7] and

layout [8] also. Image indexing using compressed transforms was

dealt by J. Berens, G. D. Finlayson and G. Qiu [9]. It uses the

standard transform encoding methods (the Karhunen-Loeve

transform, the discrete cosine transform [10] [11]. The wavelet

transform is treated in [12]. (B)

The images are very rich in the content like color, texture

and shape information present in them [2].

Retrieving

images based on color similarity usually involve comparing color histograms [11,16], color averages [4,19], BTC [20]

and other methods. Texture measures look for visual

patterns in images and how they are spatially defined [14].

The identification of specific textures in an image is

achieved pimarily by modeling texture as a two-

dimensional gray level variation, GLCM [10], vector

quantization codebooks [6], image transforms [7]. Shape

does not refer to the shape of an image but to the shape of a

particular region that is being sought out.

Shapes are often

determined by first applying segmentation or edge

detection to an image [12]. Other methods use shape filters

to identify given shapes of an image [13,14]. In some case

accurate shape detection will require human intervention

because methods like segmentation are very difficult to

completely automate [15]. Here the paper discusses shape

texture extraction using morphological operations like

erosion, dilation, top hat transform, bottom hat transform.

The block truncation coding (BTC) is applied on the

extracted shape images to obtain feature vectors of those

images which are used for CBIR.

(A)

1. Sanjay N. Talbar and Satishkumar L. Varma, "iMATCH:

   Image Matching and Retrieval for Digital Image Libraries,"

   ICETET, pp.196-201, 2009, ISBN: 978-0-7695-

   3884-6.

2. Sanjay N. Talbar and Satishkumar L. Varma,

   IRMOMENT: image indexing and retrieval by combining

   moments, IET Digest 2009, 38, DOI:10.1049/ic.2009.0148.

3. Hossein Nezamabadi-pour and Saeid Saryazdi, Object-

   Based Image Indexing and Retrieval in DCT Domain using

   Clustering Techniques, Vol. 3 JANUARY 2005 ISSN

   1307-6884.

4. M. J. Swain and D. H. Ballard, Color indexing,

   International Journal of Computer Vision, 1991, vol.7, no.1,

   pp.11-32.

5. A. K. Jain and A. Vailaya, Image retrieval using color and

   shape, Pattern Recognition, 996, vol.29, no.8, pp.1233-

   1244.

6. F. Mokhtarian and S. Abbasi, Shape similarity retrieval

   under affine transforms, Pattern Recognition, 2002, vol.

   35, pp. 31-41.

7. B. S. Manjunath and W. Y. Ma, Texture feature for

   browsing and retrieval of image data, IEEE PAMI, 1996,

   no. 18, vol. 8, pp. 837-842.

8. J. R. Smith and C. S. Li, Image classification and querying

using composite region templates, Academic Press,

Computer Vision and Understanding, 1999, vol.75, pp.165-

174.

1. J. Berens, G. D. Finlayson and G. Qiu, Image indexing

   using compressed color Histograms, IEEE Proc.-Vision

   Image Signal Process. Vol. 147, No. 4, August 2000.

2. Jose A. Lay and Ling Guan, Image Retrieval Based on

   Energy Histograms of the Low Frequency DCT

   Coefficients, IEEE 0-7803-5041-3/99, 1999.

3. Stepan Obdrzalek and Jiri Matas, Image Retrieval Using

   Local Compact DCT-based Representation. DAGM03,

   25th Pattern Recognition Symposium, September 10-12,

   2003.

4. J. Z. Wang, G. Wiederhold, O. Firschein, and X.W. Sha,

Content-Based Image Indexing and Searching Using

Daubechies Wavelets, Int'l J. Digital Libraries, vol. 1, no.

4, pp. 311-328, 1998.

(B)

[1]. Dr.H.B.Kekre, Sudeep D. Thepade, Priyadarshini

Mukherjee, Shobhit Wadhwa, Miti Kakaiya,

Satyajit Singh, Image Retrieval with Shape

Features Extracted using Gradient Operators and

Slope Magnitude Technique with BTC,

International Journal of Computer Applications,

September 2010 issue.

[2]. Dr.H.B.Kekre, Sudeep D. Thepade, Rendering

Futuristic Image Retrieval System, National Conference on Enhancements in Computer, Communication and Information Technology, EC2IT-2009, 20-21 Mar 2009, K.J.Somaiya

College of Engineering, Vidyavihar, Mumbai-77.

[3]. Image database –

http://wang.ist.psu.edu/docs/related/Image.ori g

(Last referred on 23 Sept 2008)

[4]. Dr.H.B.Kekre, Sudeep D. Thepade, Archana

Athawale, Anant Shah, Prathmesh Verlekar, Suraj

Shirke,Energy Compaction and Image Splitting for

Image Retrieval using Kekre Transform over Row

and Column Feature Vectors, International Journal

of Computer Science and Network Security (IJCSNS),Volume:10, Number 1, January 2010, (ISSN: 1738-7906) Available at www.IJCSNS.org.

[5]. Dr.H.B.Kekre, Sudeep D. Thepade, Image

Retrieval using Color-Texture Features Extracted

from Walshlet Pyramid, ICGST International

Journal on Graphics, Vision and Image Processing

(GVIP), Volume 10, Issue I, Feb.2010, pp.9- 18,

Available online

www.icgst.com/gvip/Volume10/Issue1/P11509 388

76.html

[6]. Dr.H.B.Kekre, Tanuja Sarode, Sudeep D. Thepade,

Color-Texture Feature based Image Retrieval

using DCT applied on Kekres Median Codebook,

International Journal on Imaging (IJI), Volume 2,

Number A09, Autumn 2009,pp. 55-65.

Available

online at www.ceser.res.in/iji.html

[7]. Dr.H.B.Kekre, Sudeep D. Thepade, Image

Retrieval using Non-Involutional Orthogonal Kekres Transform, International Journal of Multidisciplinary Research and Advances in

Engineering (IJMRAE), Ascent Publication House,

2009, Volume 1, No.I, pp 189-203, 2009.

Abstract

available online at www.ascent-journals.com [8]. Dr.H.B.Kekre, Sudeep D. Thepade,

Improving the

Performance of Image Retrieval using Partial

Coefficients of Transformed Image, International

Journal of Information Retrieval, Serials

Publications, Volume 2, Issue 1, 2009, pp. 72-79

[9]. Dr.H.B.Kekre, Sudeep D. Thepade, Archana

Athawale, Anant Shah, Prathmesh Verlekar, Suraj

Shirke, Performance Evaluation of Image

Retrieval using Energy Compaction and Image

Tiling over DCT Row Mean and DCT Column

Mean, Springer-International Conference on

Contours of Computing Technology (Thinkquest-

2010), Babasaheb Gawde Institute of Technology,

Mumbai, 13-14 March 2010, The paper will be

uploaded on online Springerlink.

[10]. Dr.H.B.Kekre, Tanuja K. Sarode, Sudeep D.

Thepade, Vaishali Suryavanshi,Improved Texture

Feature Based Image Retrieval using Kekres Fast

Codebook Generation Algorithm, Springer- International Conference on Contours of Computing Technology (Thinkquest-2010), Babasaheb Gawde Institute of Technology,

Mumbai, 13-14 March 2010, The paper will be

uploaded on online Springerlink.

[11]. Hirata K. and Kato T. Query by visual example

content-based image retrieval, In Proc. Of Third

International Conference on Extending Database

Technology, EDBT92, 1992, pp 56-71.

[12]. Sagarmay Deb, Yanchun Zhang, An Overview of

Content Based Image Retrieval Techniques, Technical Report, University of Southern Queensland.

[13]. Rafael C. Gonzalez, Richard E. Woods, Digital

Image Processing. Chapter 10, pg 599-607. Published by Pearson Education, Inc. 2005.

[14]. William I. Grosky, Image Retrieval – Existing

Techniques, Content-Based (CBIR) Systems

Department of Computer and Information Science,

University of Michigan-Dearborn, Dearborn, MI,

USA,http://encyclopedia.jrank.org/articles/page s/67

63/Image-Retrieval.html#ixzz0l30drFVs, referred

on 9 March 2010

[15]. Bill Green, Canny Edge Detection Tutorial, 2002.

http://www.pages.drexel.edu/~weg22/can_tut. html,

referred on 9 March 2010

[16]. John Eakins, Margaret Graham, Content Based

Image Retrieval, Chatpter 5.6, pg 36-40,

University of Northrumbia at New Castle, October

1999