Analysing Assorted Window Sizes with LBG and KPE Codebook Generation Techniques for Grayscale Image...

8/6/2019 Analysing Assorted Window Sizes with LBG and KPE Codebook Generation Techniques for Grayscale Image Colorizat…

http://slidepdf.com/reader/full/analysing-assorted-window-sizes-with-lbg-and-kpe-codebook-generation-techniques 1/5

(IJCSIS) International Journal of Computer Science and Information Security,

Vol. 9, No. 6 June 2011

Analysing Assorted Window Sizes with LBG and

KPE Codebook Generation Techniques for Grayscale

Image Colorization

Dr. H. B.KekreSr. Professor,MPSTME,

NMIMS Deemed-to-be

University,Vileparle (W),Mumbai-56, India.

Dr. Tanuja K. SarodeAsst. Professor,

Thadomal Shahani Engg.

College,Bandra (W), Mumbai-50,

India.

Sudeep D. ThepadeAsst. Professor,

MPSTME,

NMIMS Deemed-to-beUniversity,Vileparle (W),

Mumbai-56, India.

Ms. Supriya Kamoji Sr.Lecturer,

Fr.Conceicao Rodrigues

College of Engg,Bandra (W),

Mumbai-50, India.

Abstract—This paper presents use of assorted window sizes andtheir impact on colorization of grayscale images using Vector

Quantization (VQ) Code Book generation techniques. Theproblem of coloring grayscale image has no exact solution.Attempt is made to minimize the human efforts needed inmanually coloring grayscale images. Here human interaction is

only to find reference image of similar type. The job of transferring color from reference image to grayscale is done byproposed techniques. Vector quantization algorithms Linde Buzoand Gray Algorithm (LBG) and Kekre Proportionate Error

(KPE) are used to generate color palette in RGB and Kekre’s LUVcolor space. For colorization source color image is taken asreference image which is divided into non overlapping pixelwindows. Initial clusters are formed using VQ algorithms LBG

and KPE, used to generate the color palette. Grayscale imagewhich is to be colored is also divided in non overlapping pixel

windows. Every pixel window of gray image is compared withcolor palette to get the nearest color values. Best match is foundusing least mean squared error. To test the performance of thesealgorithms, color image is converted into gray scale image and the

same grayscale image is recolored back. Finally MSE of recolored

image and original image is compared. Experiment is conductedon both RGB and Kekre’s LUV color space for the different pixelwindows of size 1x2, 2x1, 2x2, 2x3, 3x2, 3x3, 1x3, 3x1, 2x4, 4x2,

1x4, 4x1. However Kekre’s LUV color space gives outstandingperformance. For different pixel windows KPE with 1x2 and LBGwith 2x1 pixel window perform well with respect to image quality.

Keywords- Colorization , Pixel Window, ColorPalette, VectorQuantization(VQ) , LBG, KPE.

I. I NTRODUCTION

Colors always provide more clear information than gray

scale digital images. Colorization is the art of adding color to a

monochrome image or movie. Colors we perceive in an objectare determined by nature of light reflected from the object. Due

to the structure of human eye, all colors are seen as variable

combinations three basic colors Red, Green, Blue (RGB). The

task of coloring a grayscale image involves assigning RGB

values to an image which varies along only the luminancevalue. Since different colors may have the same luminance but

vary in hue and saturation, the problem of coloring gray scale

needs human interaction [1].

Gray scale image is represented by only the luminance values

that can be matched between the two images. Because a singleluminance value could represent entirely different parts of an

image, the remaining values within the pixel’s neighborhood

are used to guide the matching process. Once the pixel ismatched, the color information is transferred but original

luminance value is retained [2].

The details in color image can be utilized for analysis and study

of particular image in the applications like medical

tomography, information security, image segmentation, etc.Coloring of old Black and White movies and rare images of

monuments, celebrities is one of the best applications which

give good feel and understanding.

In case of pseudo-coloring [3] where the mapping of luminance

values to color values is automatic, the choice of color map is

commonly determined by human decision. The main concept of

colorization techniques exploits textual information. The work of Welsh et al , which is inspired by the color transfer [4] and

by image analogies [5], examines the luminance values in theneighborhood of each pixel in the target image and add to its

luminance the chromatic information of a pixel from a source

image with best neighborhoods matching .This technique works

on images were differently colored regions give rise to distinct

textures otherwise, the user must specify rectangular swatchesindicating corresponding regions in the two images.

Color traits transferred to gray scale images [6] presents novel

coloring techniques where color palette is prepared using pixelwindows of some degree taken from reference coloring image.

For every window of gray scale image the palette is searched

for equivalent color values which could be used to color gray

scale window [19].

In this paper, adjacent pixels are grouped together to form a(pixel window) grid. Vector Quantization algorithms LBG andKPE are applied on different pixel window sizes 1x2, 2x1, 2x2,2x3, 3x2, 3x3, 1x3, 3x1, 2x4, 4x2, 1x4, 4x1and codebook of size512 is obtained. Vector Quantization algorithms LBG and KPEare applied. Depending on minimum Euclidean distance, LUV

134 http://sites.google.com/site/ijcsis/ISSN 1947-5500




Vol. 9, No. 6 June 2011

components of reference image are transferred to input grayimage.

II. K EKRE’S LUV COLOR SPACE [15,21]

In the proposed technique Kekre’s LUV color space is used.Where L gives luminance and U and V gives chromaticityvalues of color image. Positive values of U indicate prominenceof red components in color image and negative value of Vindicates prominence of green component. The RGB-to LUVand LUV-to-RGB conversion matrices are given in equation 1and 2 respectively.

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

V

U

L

=

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

−

110

112

111

*

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

B

G

R

(1)

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

B

G

R

=

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡ −

111

111

021

*

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

2/

6/

3/

V

U

L

(2)

III. VECTOR QUANTIZATION

Vector Quantization (VQ) [7],[8] is an efficient and lossy

technique for compression of data and has been successfully

used in various applications like an pattern recognition[11],

speech recognition and face detection[12][13],imagesegmentation[14],speech data compression [16],content based

image retrieval CBIR[17],[18] etc.

Vector Quantization can be define as a mapping function

that maps k-dimensional vector space to a finite set CB = {C1,

C2,C3, ..…., CN}. The set CB is called codebook consisting of N number of codevectors and each codevector Ci= {ci1, ci2, ci3,

……, cik } is of dimension k. The key to VQ is the good

codebook. Codebook can be generated in spatial domain by

clustering algorithms.

In color transfer phase, image is divided into non

overlapping blocks and each block then is converted to the

training vector Xi = (xi1, xi2, ……., xik ). The codebook is thensearched for the nearest codevector Cmin by computing squared

Euclidian distance as presented in equation (3) with vector Xi

with all the codevectors of the codebook CB. This method is

called exhaustive search (ES).d(Xi, Cmin) = min1≤ j≤ N{d(Xi,C j)} (3)

where d(Xi,C j) = ∑(Xip - C jp)2 It is obvious that, if the codebook size is increased to reduce thedistortion the searching time will also increase.

The following section describes the VQ codebook Generation Algorithms.

A. Linde Buzoand Gray Algorithms(LBG) [7,8]

In this algorithm centroid is first calculated by takingaverage as the first code vector for the training set. In figure1

two vectors are generated by using constant error addition to thecodevector. Euclidean distances of all the training vectors arecomputed with vectors v1 & v2 and two clusters are formed

based on closest of v1 or v2. This modus operandi is replacedfor every cluster. The shortcoming of this algorithm is that the cluster elongation is +135O to horizontal axis in two dimensionalcases resulting in inefficient clustering.

Figure1 LBG for Two dimensional case.

B. Kekre’s Proportionate Error (KPE) Algorithm [9,10]

Here to generate two vectors v1 & v2 proportionate error isadded to the codevector. Magnitude of elements of thecodevector decides the error ratio. Hereafter the procedure issame as that of LBG. While adding proportionate error a safeguard is also introduced so that neither v1 nor v2 go beyond thetraining vector space eliminating the disadvantage of the LBG.Fig. 2, shows the cluster elongation after adding proportionateerror.

Figure 2 orientation of line joining two vectors v1 and v2 after addition of

proportionate error to the centroid.

IV. PROPOSED COLORING TECHNIQUE

Since the coloring problem always requires humaninteraction. So reference image of same class and of same

feature as of input grayscale image. The color transfer

algorithm is discussed for LUV color space for different m x n

pixel grid size. The main steps of algorithm for a color transfer

are:

• Convert RGB components of source color image into

respective Kekre’s LUV color components.• Divide the image in to blocks of m x n pixels. Hence

m x n x3 dimensional training vector set

corresponding to LUV components of each pixel is

obtained. On this set LBG and KPE algorithms are

applied and color palette is generated i.e. codebook of size 512.

• The input gray image is divided in mxn blocks of

pixels. Each block (pixel window) is searched for

135 http://sites.google.com/site/ijcsis/

ISSN 1947-5500




Vol. 9, No. 6 June 2011

• nearest code vector of color palette. While searching

only luminance is compared

• Once the nearest match is obtained gray image pixel

window is replaced by LUV codevector

• The final colored image in LUV domain is then

converted into RGB plane and MSE of original color

image and recolored image are calculated.

V. R ESULTS

The algorithms discussed above are implemented using

MATLAB 7.0 on Pentium IV, 1.66GHz, 1GB RAM. To test the

performance of these algorithms we have converted color

image to grayscale image and the same gray image is recolored back. Finally MSE of original image and colored image is

compared. Five color images belonging to different classes of

size 128x128x3 are used.

Figure3 to Figure6. Shows the results of LBG and KPE for Zebra, Book, Cartoon and Face images considering same image

as reference image.

Figure7 and Figure8. Shows the results of LBG and KPE for

scenery and dog images considering different image as reference image.

6(a)

Original image

6(b)

Gray image

6(c)

1x2 Grid LBGMSE: 92

6(d)

1x2 Grid KPEMSE: 81

Figure 6 shows reconstruction of face grayscale image using similar sourceimage for pixel window 1x2.

7(a)

OriginalImage

7(b)

ReferenceImage

7(c)

GrayImage

7(d)

1x2 GridLBG

MSE 990

7(d)

1x2 Grid KPEMSE 709

Figure 7 shows reconstruction of Scenery grayscale image using

different source image.

8(a)

OriginalImage

8(b)

ReferenceImage

8(c)

Gray Image

8(d)

1x2GridLBG

MSE 303

8(e)

1x2 GridKPE MSE

285

Figure 8 shows reconstruction of Dog grayscale image using different

source image.

Various images, each of size 128x128 pixels, were

used to build the color palette, and their grayscale

equivalents were colored using color palette for various

pixel windows. The fig. 9, shows bar chart of averagemean squared error obtained across all five images with

respect to initial few pixel windows for RGB and Kekre’s

LUV color space. It is observed that, Kekre’s LUV color

space gives less MSE compared to RGB color space.

Hence in table 1 only Kekre’s LUV color space results for different images using 12 varying pixel window

sizes(1x2,2x1,2x2,2x3,3x2,3x3,1x3,3x1,2x4,4x2,1x4,4x1)

are given.

Figure 9– Average MSE across various Grid sizes for different color

spaces

3(a)

Original

image

3(b)

Gray image 3(c)

1x2 Grid LBG

MSE: 178.4

3(d)

1x2 Grid KPE

MSE: 73.85

Figure 3 shows reconstruction of Zebra grayscale image using similar

source image for pixel window 1x2

4 (a)

Original Image

4(b)

Gray Image

4(c)

1x2GridLBGMSE73.8

4(d)

1x2 Grid KPEMS53.32

Figure 4 shows reconstruction of book grayscale image using similar

source mage for pixel window 1x2

5(a)Original image

5(b)Gray image

5(c)1x2 Grid LBG

MSE: 1260

5(d)1x2 Grid KPE

MSE:1023

Figure 5 shows reconstruction of cartoon grayscale image using similar

source mage for pixel window 1x2


ISSN 1947-5500




Vol. 9, No. 6 June 2011

Table I. Shows the Results of LBG and KPE for five color images from different categories of size 128x128x3.

InputImages

VQAlg.

Grid Sizes

1x2 2x1 2x2 2x3 3x2 3x3 2x4 4x2 1x3 3x1 1x4 4x1

Image1 LBG 92.70 87.11 107.42 116.75 114.65 2302 5576 5652 107.32 106 122.29 114.

KPE 81.59 77.49 89.141 90.75 95.83 2122 5557 5663 87.32 89.09 91.41 87.3

Image2 LBG 1260 1244 1056 1493 1420 6350 6928 9595 1251 1243 1532 1392

KPE 1023 1153 1363 1233 1150 6243 6725 9359 1278 1138 1013 1093

Image3 LBG 178.4 147 440.90 721.73 675.15 2003 2486 4236 373 286 610.66 465.7

KPE 73.85 76.12 225.10 461.69 441.62 1823 2264 4663 143.5 142 273.38 226.0

Image4 LBG 73.89 76.64 107.12 123.82 131.86 2340 3400 3664 90 97 116.92 128.9

KPE 53.32 52.73 79.094 103.43 95.439 2813 3371 3701 65.3 65 73.93 76.09

Image5 LBG 1203 1244 1244 1406 1388 6340 7833 7916 1246 1240 1274 1266

KPE 1178 1174 1182 1414 1399 6384 7935 7968 1193 1175 1209 1212

Average LBG 561.5 559.7 591 772.26 745.9 3867 5244 6212 613 594 731 673

Average KPE 481.9 506.6 587.8 660.5 636.3 3877 5170 6270 553.4 521.8 532.1 538.8

From the data given in table1, it is seen that the performance

gradually decreases as the pixel window size increases.Further

MSE for unidirectional pixel window is less compared to

bidirectional. Pixel window sizes 1x2 and 2x1 are showing

better results as compared to large pixel window sizes.Fig.10, shows the comparison of average mean sqared error

obtained across all images on Kekre’s LUV color space for topfive pixel windows. It can be seen from the chart , KPE

performs well with respect to LBG. Also performance

deteriorates as pixel window size increases and becomes

bidirectinal.

Figure 10 Average MSE across various Grid sizes

VI. CONCLUSION

In this paper, the idea of colorization of grayscale imagesusing VQ codebook generation techniques is presented usingtwo famous codebook generation algorithms alias LBG andKPE. For both the algorithms 12 assorted pixel window sizes areconsidered for preparing the color palettes. As quality of colorization is subjective to source color image and grayscale to

be colorized image, the grayscale version of 5 color images arerecolored using total 48 variations of proposed techniques with 2color spaces (RGB and Kekre’s LUV), 12 pixel window sizes

and 2 codebook generation techniques( LBG and KPE). Thecomparison of original color image and recolored image hasshown that Kekre’s LUV color space outperforms RGB color space. Further, it can be observed from results that unidirectional

pixel windows gives better colorization than bidirectional pixelwindow sizes. The KPE performs better than LBG for colorization. In all the best performance is shown by KPE with1x2 window size in Kekre’s LUV color space.

R EFERENCES

[1] V. Karthikeyani, K. Duraisamy, Mr.P.Kamalkakkannan, " Conversionof grayscale image to color image with and without texture synthesis",IJCSNS International journal of Computer science and network security, Vol.7 No.4 April 2007.

[2] E.Reinhard, M. Ashikhmin, B. Gooch and P Shirley, “Colour Transfer between images”, IEEE Transactions on Computer Graphics andApplications 21, 5, pp. 34-41.

[3] Rafael C. Gonzalez & Paul Wintz, “ Digital Image Processing”,Addison Wesley Publications, May 1987.

[4] A. Hertzmann, C. E Jacobs, N. Oliver, B. Curless and D.H. Salesin,“image Anologies”, in the proceedings of ACM SIGGRAPH 2002, pp.341-346.

[5] G. Di Blassi, and R. D. Reforgiato, “Fast colourization of grayimages”, In proceedings of Eurographics Italian Chapte, 2003.

[6] H.B.Kekre, Sudeep. D. Thepade, “Color traits transfer to gray scaleimages”, in Proc of IEEE International conference on Emerging Trendsin Engineering and Technology, ICETET 2008 Raisoni College of Engg, Nagpur.

[7] R. M. Gray, "Vector quantization", IEEE ASSP Mag., pp. 4-29,

Apr11984.[8] Y. Linde, A. Buzo, and R. M. Gray, "An algorithm for vector quantizer

design," IEEE Trans.Commun., vol. COM-28, no. 1, pp. 8495, 1980.

[9] H. B. Kekre, Tanuja K. Sarode, "New Fast Improved Codebook Generation Algorithm for Color Images using Vector Quantization,"International Journal of Engineering and Technology, vol.1, No.1, pp.67-77, September 2008.

[10] H. B. Kekre, Tanuja K. Sarode, "An Efficient Fast Algorithm toGenerate Codebook for Vector Quantization," First InternationalConference on Emerging Trends in Engineering and Technology,ICETET-2008, held at Raisoni College of Engineering, Nagpur, India,July 2008, Available at online IEEE Xplore.

[11] Ahmed A. Abdelwahab, Nora S. Muharram, "A Fast Codebook DesignAlgorithm Based on a Fuzzy Clustering Methodology", InternationalJournal of Image and Graphics, vol. 7, no. 2 pp. 291302, 2007.

[12] Chin-Chen Chang, Wen-Chuan Wu, "Fast Planar-Oriented Ripple

Search Algorithm for Hyperspace VQ Codebook", IEEE Transactionon image processing, vol 16, no. 6, June 2007.

[13] C. Garcia and G. Tziritas, "Face detection using quantized skin color regions merging and wavelet packet analysis," IEEE Trans.Multimedia, vol. 1, no. 3, pp. 264-277, Sep. 1999.

[14] H. B. Kekre, Tanuja K. Sarode, Bhakti Raul, "Color ImageSegmentation using Kekre's Fast Codebook Generation AlgorithmBased on Energy Ordering Concept", ACM International Conferenceon Advances in Computing, Communication and Control (ICAC3-2009), 23-24 Jan 2009, Fr. Conceicao Rodrigous College of Engg.,

Mumbai. Available on online ACM portal .


ISSN 1947-5500




Vol. 9, No. 6 June 2011[15] Dr.H.B. Krkre, Sudeep D. Thepade, “Image Blending in Vista Creation

using Kekre’s LUV Color Space”, In Proc. Off PIT-IEEE Colloquium,Mumbai, Feb 4-5,2008.

[16] H. B. Kekre, Tanuja K. Sarode, "Speech Data Compression usingVector Quantization", WASET International Journal of Computer andInformation Science and Engineering (IJCISE), vol. 2, No. 4, 251254,Fall 2008. available: http://www.waset.org/ijcise.

[17] H. B. Kekre, Ms. Tanuja K. Sarode, Sudeep D. Thepade, "ImageRetrieval using Color-Texture Features from DCT on VQ Codevectorsobtained by Kekre's Fast Codebook Generation", ICGST-InternationalJournal on Graphics, Vision and Image Processing (GVIP),Volume 9,Issue 5, pp.: 1-8, September 2009. Available online at http://www.icgst.com/gvip/Volume9/Issue5/P1150921752.html.

[18] H.B.Kekre, Tanuja K. Sarode, Sudeep D. Thepade, "Color-TextureFeature based Image Retrieval using DCT applied on Kekre's MedianCodebook", International Journal on Imaging (IJI), Available online at www.ceser.res.in/iji.html.

[19] Dr. H. B. Kekre, Sudeep D. Thepade, Nikita Bhandari, “Colorization of Gereyscale images using Kekre’s Bioorthogonal Color Spaces andKekre’s Fast Codebook Generation “,CSC Advances in MultimediaAn international journal (AMU), volume 1, Issue 3,pp.49-58, Available

at www.cscjournals.org/csc/manuscript/journals/AMIJ/volume1/Issue3/A

MU-13.pdf.

[20] Dr. H. B. Kekre, Sudeep D. Thepade,Adib Parkar, “A Comparison of Harr Wavelets and Kekre’s Wavelets for Storing Color Information in

a Greyscale Images”, International Journal of Computer Applications(IJCA), Volume 1, Number 11, December 2010,pp 32-38.

Available at www.ijcaonline.org/archives/volume11/number11/1625-2186 .

[21] Dr. H. B. Kekre, Sudeep D. Thepade,Archana Athawale, Adib Parkar,“Using Assorted Color Spaces and pixel window sizes for Colorizationof Grayscale images’,ACM International Conferences and workshopson emerging Trends in Technology(ICWET 2010), Thakur College of Engg. And Tech.,Mumbai,26-27 Feb 2010.

[22] H. B. Krekre,Sudeep Thepade, Adib Parkar, “A comparison of Kekre’sFast Search and Exhaustive Search for various grid sizes used for coloring a Grayscale Image” Second International conference on signalAcquisition and Processing, (ICSAP2010), IACSIT,Banglore,pp.53-57,9-10 Feb 2010.

Author Biographies

Dr. H. B. Kekre has received B.E. (Hons.) in Telecomm. Engineering. from

Jabalpur University in 1958, M.Tech(Industrial Electronics) from IIT Bombay in

1960, M.S.Engg. (Electrical Engg.) from

University of Ottawa in 1965 and Ph.D.

(System Identification) from IIT Bombayin 1970 He has worked as Faculty of

Electrical Engg. and then HOD Computer

Science and Engg. at IIT Bombay. For 13years he was working as a professor and head

in the Department of Computer Engg. at

Thadomal Shahani Engineering. College, Mumbai. Now he is Senior Professor

at MPSTME, SVKM’s NMIMS. He has guided 17 Ph.Ds, more than 100M.E./M.Tech and several B.E./ B.Tech projects. His areas of interest are Digital

Signal processing, Image Processing and Computer Networking. He has more

than 270 papers in National / International Conferences and Journals to hiscredit. He was Senior Member of IEEE. Presently He is Fellow of IETE and

Life Member of ISTE Recently 11 students working under his guidance have

received best paper awards. Two of his students have been awarded Ph. D. from NMIMS University. Currently he is guiding ten Ph.D. students.

Dr. Tanuja K. Sarode has Received Bsc.(Mathematics) from MumbaiUniversity in 1996, Bsc.Tech.(Computer

Technology) from Mumbai University in 1999,

M.E. (Computer Engineering) degree from Mumbai

University in 2004, Ph.D. from Mukesh Patel

School of Technology, Management andEngineering, SVKM’s NMIMS University, Vile-

Parle (W), Mumbai, INDIA. She has more than 12

years of experience in teaching. Currently working

as Assistant Professor in Dept. of Computer

Engineering at Thadomal Shahani EngineeringCollege, Mumbai. Engineering, SVKM’s NMIMS University, Vile-Parle (W),

Mumbai, INDIA. She has more than 12 years of experience in teaching.

Currently working as Assistant Professor in Dept. of Computer Engineering atThadomal Shahani Engineering College, Mumbai. She is life member of IETE,

member of International Association of Engineers (IAENG) and International

Association of Computer Science and Information Technology (IACSIT),

Singapore. Her areas of interest are Image Processing, Signal Processing and

Computer Graphics. She has 90 papers in National /InternationalConferences/journal to her credit.

Sudeep D. Thepade has Received B.E.(Computer) degree from North

Maharashtra University with Distinction in2003. M.E. in Computer Engineering from

University of Mumbai in 2008 withDistinction, currently submitted thesis for

Ph.D. at SVKM’s NMIMS, Mumbai. He has

more than 08 years of experience in

teaching and industry. He was Lecturer inDept. of Information Technology at

Thadomal Shahani Engineering College,

Bandra(w), Mumbai for nearly 04 years.

Currently working as Associate Professor in

Computer Engineering at Mukesh Patel School of Technology Managementand Engineering, SVKM’s NMIMS, Vile Parle(w), Mumbai, INDIA. He is

member of International Association of Engineers (IAENG) and International

Association of Computer Science and Information Technology (IACSIT),

Singapore. He is member of International Advisory Committee for manyInternational Conferences. He is reviewer for various International Journals.

His areas of interest are Image Processing Applications, Biometric

Identification. He has about 110 papers in National/International

Conferences/Journals to his credit with a Best Paper Award at InternationalConference SSPCCIN-2008, Second Best Paper Award at ThinkQuest-2009

National Level paper presentation competition for faculty, Best paper award atSpringer international conference ICCCT-2010 and second best research project

award at ‘Manshodhan-2010’.

Supriya Kamoji has received B.E. in Electronics and Communication

Engineering with Distinction from Karnataka

University in 2001. Currently pursuing M.E. fromThadomal Shahani College of Engineering,

Mumbai, India. She has more than 8years of

teaching experience. Currently working as an

Senior Lecturer in Fr.Conceicao Rodrigues

College of Engineering. Mumbai, India. She is alife time member of Indian society of Technical

Education (ISTE). Her areas of interest are Image

Processing, Computer Organization and Architecture and Distributed

Computing.


ISSN 1947-5500

Analysing Assorted Window Sizes with LBG and KPE Codebook Generation Techniques for Grayscale Image...

Documents

Transcript of Analysing Assorted Window Sizes with LBG and KPE Codebook Generation Techniques for Grayscale Image...