8/15/2019 Project Presentation Image Processing
1/14
Under the guidance of..Dr.k.GiribabuHead of the department
Electronics & Communication Engineering
By.... T. ounica. !ukanya
". Go#thami. Bha$ya
Content Based Image Retrieval(CBIR)
8/15/2019 Project Presentation Image Processing
2/14
CONTENTS INTRODUCTION
IMAGE RETRIEVA ARC!ITECTURE
"!# CBIR$$$$$
CBIR MODE
!O" CBIR "OR%S$$$$
&EATURES O& IMAGE
A''ICATIONS
CONCUSION
8/15/2019 Project Presentation Image Processing
3/14
RODUCTION
ontent based image retrieval, a technique which uses visual
contents from large scale image database according to users
query, has been an active and fast advancing research area
since the 1990s.
n our project we concentrated on regionhistogram features toetrieve the images according to an e!ample query image
upplied by the user.
e ha$e proposed a CB% based image retrie$al system'#hich analyses innate properties of an image such as' thete(ture' and histogram for e)cient and meaningfulimage retrie$al.
8/15/2019 Project Presentation Image Processing
4/14
IMAGE RETRIEVA ARC!ITECTURE
%mage
collection
"isual
features
Te(t
annotations
ulti dimensional inde(ing
*uery processing
*uery interface
+eaturee(traction
user
8/15/2019 Project Presentation Image Processing
5/14
MET!ODS INVOVED
Tet *ased image retrieval (TBIR)
Content *ased image retrieval (CBIR)
8/15/2019 Project Presentation Image Processing
6/14
"!# CBIR$$$$$$
Contentbased means that the search analyses the contents of the image not the metadata such as "eywords, labels or
tags associated with the images.
It is also "nown as query by image content #$%IC& and
contentbased visual information retrieval #C%'I(&
)eatures such as colour , te!ture, shape and spatial are retrieved
automatically
*imilarities of the images are based on distance between the features
+o need of domain e!perts
escription of image in te!t form doesn-t required
8/15/2019 Project Presentation Image Processing
7/14
CBIR MODE
Fig: Block diagram of CBIR system
8/15/2019 Project Presentation Image Processing
8/14
!O" CBIR "OR%S$$$$$$
8/15/2019 Project Presentation Image Processing
9/14
&EATURES O& IMAGE
Colo+r
S,a-e
Tet+re
8/15/2019 Project Presentation Image Processing
10/14
COOR
Color similarity is achieved by computing a color histogram for each image that identifies the
proportion of pi!els within an image holding specific values
#that humans e!press as colors&.
!amining images based on the colors they contain
is one of the most widely used techniques because it
does not depend on image si/e or orientation.
Color searches will usually involve comparing
color histograms, though this is not the only technique in
practice.
http://en.wikipedia.org/wiki/Color_histogramhttp://en.wikipedia.org/wiki/Color_histogram
8/15/2019 Project Presentation Image Processing
11/14
S!A'E
*hape does not refer to the shape of an image but to the
shape of a particular region that is being sought out.
*hapes will often be determined first applying segmentatio
or edge detection to an image.
ther methods li"e use shape filters to identify given shapes of an image.
http://en.wikipedia.org/wiki/Segmentation_(image_processing)http://en.wikipedia.org/wiki/Edge_detectionhttp://en.wikipedia.org/wiki/Edge_detectionhttp://en.wikipedia.org/wiki/Segmentation_(image_processing)
8/15/2019 Project Presentation Image Processing
12/14
TE.TURE
e!ture measures loo" for visual patterns in images and
how they are spatially defined.
hese sets not only define the te!ture, but also where in
the image the te!ture is located.
e!ture is a difficult concept to represent. he identificatio of specific te!tures in an image is achieved primarily by
modeling te!ture as a twodimensional gray level variation.
8/15/2019 Project Presentation Image Processing
13/14
ICATIONS
edi/al A--li/ations
ood e(ample of the po#er of image retrie$al is that of dist
dicine
e* A--li/ations
d to retrieve digital images from the large database search for oneeci-c image
arch for a picture to go #ith a broad story or search to illustdocument.
8/15/2019 Project Presentation Image Processing
14/14
T,an0 +