International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 12 | Dec-2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 828
A Survey on Content Based Image Retrieval for Reducing Semantic Gap
T.RajaSenbagam
1
, Dr.R.Shanmugalakshmi
2
1
Assistant Professor, Department of CSE & Government College of Technology, Coimbatore-13
2
Professor & Head, Department of EEE & Government College of Engineering, Salem Tamil Nadu, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Image retrieval is the process of searching and
retrieving images from large image database by usinga query
image. Now there is an increasing volume of digital images
available in the World Wide Web produced by scientific,
educational, medical, industrial, and other applications are
easily available to users. But it is a challenging task for the
user to retrieve an image from this huge volume of image
database by manual annotation method or otherwise calledas
text based retrieval method. Hence we move on to a newimage
retrieval method called content based image retrieval
(CBIR).By combining the low level features and high level
semantic feature an image can be retrieved in CBIR. Content
based image retrieval uses the visual content of the image to
retrieve the image from large image database. But reducing
the semantic gap between the low level visual features and the
high level image semantics is a challenging task in content
based image retrieval. Here in this paper we have provided a
comparative study of various techniques of CBIR and the
various techniques to reduce the semantic gap between the
query image and the retrieved image.
Key Words: Content Based Image Retrieval, Semantic
Gap, Low level feature, High level feature, Relevance
Feedback, Text Based Image retrieval
1. INTRODUCTION
Content Based Image Retrieval (CBIR) is a searching
technique that can be used for retrieving most similar
images from the large collection of database and it is also
called as Query By Image Content (QBIC) [1]. The term
content refers to colors, shapes, textures or any other
information that can be derived from the image itself. Hence
the recognition and retrieval of information in CBIR is based
on content of images only instead of metadata such as
keywords, tags or descriptions associated with the images
[2].
Types of Image Retrieval
Image retrieval is of two categories:
1. Text-Based Image Retrieval
2. Content-Based Image Retrieval
Text-Based Image Retrieval
Text-based image retrieval is a traditional technique to
retrieve an image from database. It uses the keyword
annotation to retrieve the image. It has several advantages
such as ease of retrieval, adaptable for small databases and
keyword usage for representing more images. But it is still
time consuming, incomplete and not standardized.
Content-Based Image Retrieval
CBIR uses the visual contents of the image to retrieve digital
image from the large set of databases. The visual contents of
an image are the low level features such as color, shape,
texture and spatial layout. The main advantage of CBIR is
having more accuracy when compares with text based
image retrieval and reduction of dissimilarity between the
semantic content of images. But the image retrieval
procedure is complex and slower when compared to text
based system.
Fig:1.1 CBIR System
The Overview of CBIR system is shown in fig. 1.1. Consider
this CBIR system, in which the images are captured by the
digital camera and the common features are extracted from
the captured image which will be stored in the image
database. The features are extracted from various images in
the image database by applying feature extraction
algorithms and the feature vector values of different images
are stored in the feature vector Database. In order to retrieve
a particular image from this huge vector database a query
image is given as input to the system , the system will then
extract the features such as color, texture and shape from the
given query image and compute the feature vector value of
the query image. The system will match the feature vector of
query image with the image in the database and retrieve the