Professional Documents
Culture Documents
Wavelet Transform Based Image Retrieval System by Using Interactive Genetic Algorithm1
Wavelet Transform Based Image Retrieval System by Using Interactive Genetic Algorithm1
1
Department of ECE, JNTU College of Engineering,Anatapuramu,chestialtaff@gmail.com
2
Principal, Narasaraopet Institute of Technology,Narasaraopet,dv2002in@yahoo.co.in
3
Assistant Professor, JNTU College of Engineering,Anantapuramu,aruna_mastani@yahoo.com
Abstract: CBIR has become an active and fast- for better retrieval of image according to the
advancing research area in image retrieval. The application. Approximation of the user’s
primary goal of the CBIR system is to construct information need for the following search in the
meaningful descriptions of physical attributes from
image database, involving user’s interaction is
images to facilitate efficient and effective retrieval. In
necessary for a CBIR system.
order to increase the accuracy of image retrieval, a
Chin-Chin Lai et.al.[23] have proposed an
content-based image retrieval system based on
interactive genetic algorithm (IGA) is proposed. The
interactive genetic algorithm (IGA) to reduce the
retrieval process is divided into two stages. In the gap between the retrieval results and the users’
query stage, the feature descriptors of a query image expectation .They have used Color attributes like
were extracted and then used to evaluate the the mean value, standard deviation, and image
similarity between the query image and those images bitmap .They have also used texture features like
in the database. In the evolution stage, the most the entropy based on the gray level co-occurrence
relevant images were retrieved by using the IGA.
matrix and the edge histogram. Sung-Bae Cho and
IGA is employed to help the users identify the images
Joo-Young Lee [24] have proposed A Human-
that are most satisfied to the users’ need. The
Oriented Image Retrieval System that extracts the
experimental evaluation of the system is based on a
10000 WANG color image database. Experimental
features from images by wavelet transform and
results demonstrate the feasibility of the proposed provides a user-friendly means to retrieve an image
approach. from a large database when the user cannot clearly
define what the image must be. Linying Jiang
Keywords: Discrete Wavelet Transform, gray level co- et.al[25] have proposed CBIR algorithm oriented
occurrence matrix, color covariance matrix, interactive by Users’ Experience in order to improve the
genetic algorithm
storage efficiency, retrieval speed and accuracy of
the existing CBIR algorithm as well as to improve
I. INTRODUCTION the quality of user experience.
HSV is an intuitive color space in the sense that
Content based features are mainly divided into two
each component contributes directly to visual
domains; Common visual features and field
perception, and it is common for image retrieval
specific visual features like face recognition, task
systems. Hue is used to distinguish colors, whereas
dependent applications etc. On the other hand, high
saturation gives a measure of the percentage of
level features include semantic based image
white light added to a pure color. Value refers to
retrieval computed from text description or by
the perceived light intensity. The important
complex algorithms of visual features [1,2]. The
advantages of HSV color space are as follows:
mixture of these content based features is required
good compatibility with human intuition,
separability of chromatic and achromatic either a low pass frequency operation or
components, and possibility of preferring one high pass frequency operation to the rows,
component to other. The conversion formula’s are and the second letter refers to the filter
as following: applied to the columns. The lowest
resolution level LL consists of the
1
⌊(𝑅 − 𝐺 ) + ( 𝑅 − 𝐵)⌋ approximation part of the original image.
𝐻 = cos −1 ( 2 )
√( 𝑅 − 𝐺 )2 + (𝑅 − 𝐵)(𝐺 − 𝐵) The remaining three resolution levels
consist of the detail parts and give the
3
𝑆 =1− [min( 𝑅, 𝐺, 𝐵)] vertical high (LH), horizontal high (HL) and
𝑅+𝐺+𝐵
high (HH) frequencies. Figure 1 shows
𝑅+𝐺+𝐵
𝑉= three-level wavelet decomposition of an
3
image.
Texture is an important attribute that refers to
innate surface properties of an object and their
relationship to the surrounding environment. If we
could choose appropriate texture descriptors, the
performance of the CBIR should be improved. We
use a gray level co-occurrence matrix (GLCM),
which is a simple and effective method for
representing texture.
IV. CONCLUSION
V. REFERENCES