cbir using colour and texture features

25
Presented by; Jasmy Elizebeth Jose S7 E C E, Roll No:27 SAINTGITS College of Engineering 1 CONTENT BASED IMAGE RETRIEVAL BY USING COLOUR AND TEXTURE FEATURES. GUIDED BY: Er. Hanna Mathew

Upload: jasmy-elizebeth-jose

Post on 21-Jan-2017

60 views

Category:

Engineering


2 download

TRANSCRIPT

Page 1: CBIR USING COLOUR AND TEXTURE FEATURES

Presented by;

Jasmy Elizebeth Jose

S7 E C E, Roll No:27

SAINTGITS College of Engineering

1

CONTENT BASED IMAGE RETRIEVAL BY USING

COLOUR AND TEXTURE FEATURES.

GUIDED BY: Er. Hanna Mathew

Page 2: CBIR USING COLOUR AND TEXTURE FEATURES

CONTENTS:

Introduction Why we go for CBIR ? Colour feature extraction Texture feature extraction CBIR Advantages Challenges Applications of CBIR Limitations and future scope. Conclusion References

2

Page 3: CBIR USING COLOUR AND TEXTURE FEATURES

INTRODUCTION; Content-based image retrieval (CBIR) is the

application of computer vision/image processing to the image retrieval problem, that is, the problem of searching for digital images in large databases depending upon the contents of images known as features .

Features can be of two types; - low level

- high level. Image retrieval can be of two types - Text based - Content based The term CBIR seems to have originated in 1992, when

it was used by T. Kato. 3

Page 4: CBIR USING COLOUR AND TEXTURE FEATURES

CONTENT BASED IMAGE RETRIEVAL. Why CBIR????

Digital image database growing rapidly in sizeProfessional needs – Logo SearchDifficulty in locating images on the web Limitations inherent in metadata based

systems large range of possible uses for efficient

image retrieval.Possibility of missing images which use

different synonyms 4

Page 5: CBIR USING COLOUR AND TEXTURE FEATURES

COLOUR AS A FEATURE: first and most straight forward visual feature for

image retrieval. robust and simple to represent. colour system is independent of the underlying image

device. exhibit perceptual uniformity.

5

Page 6: CBIR USING COLOUR AND TEXTURE FEATURES

COLOUR FEATURE EXTRACTION.

6

START

RGB HSV

QUANTIFICATION

FEATURE VALUE COUNT

HSV

EUCLIDEAN DISTANCE

STOP

QUANTIFICATION

H=

S= V =

EUCLIDEAN DISTANCE

D= (Ai- Bi)²n

i =1

Page 7: CBIR USING COLOUR AND TEXTURE FEATURES

TEXTURE AS A FEATURE:Texture measures look for visual patterns in

images and how they are spatially defined.

Refers to properties that represents the surface or structure of an object

Features which represents texture is also called tones based on pixel intensity.

Texture is a difficult concept to represent. The identification of specific textures in an image is achieved primarily by modeling texture as a two-dimensional gray level variation.

7

Page 8: CBIR USING COLOUR AND TEXTURE FEATURES

TEXTURE FEATURE EXTRACTION.

8

START

RGB GREY SCALE

QUANTIFICATION

FEATURE VALUE COUNT

INTERNAL NORMALIZATION

STOP

QUANTIFICATION

Page 9: CBIR USING COLOUR AND TEXTURE FEATURES

RGB to GREY SCALE CONVERSION

Y=0.29xR+0.58xG+0.114xB Y=Grey scale value. R,G,B = red, green, blue colour. FEATURE MATRIX OF AN IMAGE

9

Hi=[hi,1,hi,2,……………hi,N].

INTERNAL NORMALISATION

hij = ( hij – m j )

j

Page 10: CBIR USING COLOUR AND TEXTURE FEATURES

CBIR BLOCK DIAGRAM;

10

DATABASE IMAGES

FEATURE EXTRACTION

QUERY FEATURES

FEATURE DATABAS

E

FEATURE SIMILARIT

Y MEASURE MATCHIN

G AND INDEXING

RETRIEVED IMAGES

QUERY IMAGE

Page 11: CBIR USING COLOUR AND TEXTURE FEATURES

CBIR USING COLOUR & TEXTURE FUSED FEATURES. Query is given from the user Colour and texture features are extracted

calculated and these stored in a matrix called feature matrix.

Feature vector also formed for the images present in the database

Euclidean distance is calculated between the feature vector of query image & database image

sorted the distance in ascending order and top K images are displayed on the screen. 11

Page 12: CBIR USING COLOUR AND TEXTURE FEATURES

ADVANTAGES OF CBIR:

Search for one specific image. General browsing to make an interactive choice. Search for a picture to go with a broad story or

search to illustrate a document. Good option for wide databases. Reduces human labour for interpretation.

12

Page 13: CBIR USING COLOUR AND TEXTURE FEATURES

CHALLENGES:

Semantic gapThe semantic gap is the lack of coincidence between

the information that one can extract from the visual data .

Huge amount of objects to search among.

13

Page 14: CBIR USING COLOUR AND TEXTURE FEATURES

APPLICATION AREAS

Architectural and engineering designArt collectionsCrime preventionGeographical information and remote sensing systemsIntellectual propertyMedical diagnosisMilitaryPhotograph archivesFace FindingTextiles Industry 14

Page 15: CBIR USING COLOUR AND TEXTURE FEATURES

LIMITATIONS:

Not having an universally acceptable algorithmic means of characterizing human vision in the context of image understanding.

building up on prior work or exploring novel directions is difficult.

Time for retrieval is more for databases with more than thousands of images.

FUTURE SCOPE: CBIR Using colour - texture – shape fused low level

features. CBIR Using high level features , multidimensional

features. 15

Page 16: CBIR USING COLOUR AND TEXTURE FEATURES

CONCLUSION:

CBIR based on color and texture features with that of the color and texture fused features, it is observed that CBIR based on color and texture fused features provides better results .

Other low level feature such as shape will be fused to make the image retrieval more efficient in future.

16

Page 17: CBIR USING COLOUR AND TEXTURE FEATURES

REFERENCES:[1] Aparna G Nair,C S Gode“Enhancement of image retrieval by using

colour, texture and shape features”,International conference on Electronic systems,signal processing and computing technologiesDOI 10.1109/ICESC.2014.48

[2] Swapnalini Pattanaik,D G Bhalke.”Beginners to content based image retrieval”.IJSRET@2012

[3] S Jayaraman, S Esakirajan,T Veerakumar.”Digital Image processing” McGraw Hill Education Private Limited,2013.

[4] Jun Yue, Zhenbo Li, Lu Liu, Zetian Fu. “Content based image retrieval using color and texture fused features”, 1121-1127, 2011, Elsevier.

[5] Jagadeesh Pujari, Pushpalata S. N., “Content based image retrieval using color and shape descriptors”,2010 IEEE.

[6] www.wikipidea.org/CBIR/applications and research areas[7] http://www.research.ibm.com/topics/popups/deep/manage/ html/qbic.html 17

Page 18: CBIR USING COLOUR AND TEXTURE FEATURES

EXISTING COMMERCIAL CBIR SYSTEMS EXAMPLES.- QBIC-IBM

18

SOURCE:http://www.research.ibm.com/topics/popups/deep/manage/html/qbic.html

Page 19: CBIR USING COLOUR AND TEXTURE FEATURES

QUERY SEARCH BASED ON SHAPE IN QBIC

19

SOURCE: http://www.research.ibm.com/topics/popups/deep/manage/html/qbic.html

Page 20: CBIR USING COLOUR AND TEXTURE FEATURES

QUERY SEARCH BASED ON TEXTURE - QBIC

20http://www.research.ibm.com/topics/popups/deep/manage/html/qbic.html

Page 21: CBIR USING COLOUR AND TEXTURE FEATURES

INPUT IMAGE FOR RETRIEVAL – QUERY IMAGE.

21Source: Apurva N Ganar,C S Gode,Sachin M Jambhulkar,”Enhancement of Image Retrieval by using colour, texture and shape Features,2014 international conference on electronic system.ICESC.2014.48

Page 22: CBIR USING COLOUR AND TEXTURE FEATURES

CBIR USING COLOUR FEATURE.

22Source: Apurva N Ganar,C S Gode,Sachin M Jambhulkar,”Enhancement of Image Retrieval by using colour, texture and shape Features,2014 international conference on electronic system.ICESC.2014.48

Page 23: CBIR USING COLOUR AND TEXTURE FEATURES

CBIR USING TEXTURE FEATURES.

23Source: Apurva N Ganar,C S Gode,Sachin M Jambhulkar,”Enhancement of Image Retrieval by using colour, texture and shape Features,2014 international conference on electronic system.ICESC.2014.48

Page 24: CBIR USING COLOUR AND TEXTURE FEATURES

CBIR USING COLOUR & TEXTURE FUSED FEATURES

24Source: Apurva N Ganar,C S Gode,Sachin M Jambhulkar, ”Enhancement of Image Retrieval by using colour, texture and shape Features”,2014 international conference on electronic system.ICESC.2014.48

Page 25: CBIR USING COLOUR AND TEXTURE FEATURES

25