International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014
ISSN 2319 - 4847
Content based Image Retrieval System using
Sketches and Colored Images with Clustering
Prof. T. D. Khadtare1, Sagar Ghan2 , Ekta Konde3 , Rutuja Inamke4
Abstract
The content based image retrieval (CBIR) is one of the most popular, rising research areas of the digital image processing. Most
of the available image search tools, such as Google Images and Yahoo! Image search, are based on textual annotation of images.
In these tools, images are manually annotated with keywords and then retrieved using text-based search methods. Therefore the
performances of these systems are not satisfactory. The goal of CBIR is to extract visual content of an image automatically, like
colour, texture, and shape. Our project aims to design and develop a CBIR system, which is based on sketch and coloured images.
With the help of the existing methods, we will describe a possible solution of how to design and implement a CBIR system on the
basis of colour, shape and texture as parameters, and also handle the informational gap between a sketch and a coloured image,
making an opportunity for the efficient search hereby. It is constructed after special sequence of pre-processing steps so that the
transformed full colour image and the input image can be compared. Overall, the results show that the sketch based and colour
image based system allows users an intuitive access to search-tools. The CBIR technology can be used in several applications
such as digital libraries, crime prevention, photo sharing sites, etc. Such a system has great value in apprehending suspects and
indentifying victims in forensics and law enforcement.
Keywords: CBIR, colour, shape, texture.
1. Introduction
Content Based Image Retrieval (CBIR) is an automatic process to search relevant images based on user input. The input
could be parameters, sketches or example images. A typical CBIR process first extracts the image features and store them
efficiently. Then it compares with images from the database and returns the results.
Feature extraction and similarity measure are very dependent on the features used. In each feature, there would be more
than one representation. Among these representations, histogram is the most commonly used technique to describe
features.
Although content based methods are efficient, they cannot always match user's expectation. Relevance Feedback (RF)
techniques are used to adjust the query by user's feedback. RF is an interactive process to improve the retrieval accuracy by
a few iterations. RF algorithms are dependent on feature representations.
2. Architecture Diagram
Figure 1: System Architecture
The Figure 1 shows the architecture of CBIR system. In this system query image is taken in the form of a sketch image or a
coloured image. The feature vector consisting of colour, shape and texture is calculated for database and query images.
After the process of matching the feature vectors of query image and database image, cluster of relevant images is then
displayed as output. Sequential accessing of database image is done.
3. System Implementation and results
Working of the system with input and output images:- In the system sketch as well as colored Input image is taken and
relevant output is shown as follows.
3.1 Input given as colored image-
Volume 3, Issue 5, May 2014
Page 114
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014
ISSN 2319 - 4847
Figure 2: Coloured image as Input
This is the view of working of CBIR system with input as a colored image and the relevant searches. Thus the search with
highest rate of accuracy is seen on top. Accuracy is measured in terms of distance which is calculated using Euclidean
algorithm. The pixels of input image are read and using RGB model colours of each pixel are extracted. Based on relevancy
of images, clusters are formed for efficient retrieval.
3.2 Input given as a sketch:-
Figure 3: Sketched Image as Input
Volume 3, Issue 5, May 2014
Page 115
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014
ISSN 2319 - 4847
This is the view of working of CBIR system with input as an sketch image and the relevant searches. Thus the search with
highest rate of accuracy is seen on top. Accuracy is measured in terms of distance which is calculated using Euclidean
algorithm. For a sketch image feature extraction, edges are extracted. Texture of the image is also considered for evaluation
of sketch image. Based on relevancy of images, clusters are formed for efficient retrieval.
3.3 PRECISION & RECALL GRAPH DIAGRAM:-
Figure 4: Recall and Precision Graph
These are the approximately estimated results for precision graph calculated from the formula as follows:-
4. Conclusion
We are designing a system which will retrieve relevant images related to the query input image and based on the features
such as texture, colour and sketch. The searching is done based on three types of parameters sketched image, colored
image and both(sketch as well as colour). For further improving the accuracy user feedback can be used.
5. Acknowledgement
We take this opportunity to thank Prof. T. D. Khadtare. This research paper cannot be considered complete without
mentioning his name. We wish to express true sense of gratitude towards his valuable contribution. We are grateful to his
constant encouragement and guidance in the fulfillment of this activity.
References
[1] J. Wang and G. Wiederhold, "SIMPLIcity: Semantics-Sensitive Integrated Matching for Picture Libraries," IEEE
Transactions On Pattern Analysis And Machine Intelligence, vol. 23, no. 8, pp. 1-17, September 2008.
[2] M. Swain and D. Ballard, "Color indexing," International Journal of Computer Vision, vol.. 7, no.1, pp. 11-32, 2008.
[3] K. Hirata and T. Kato, "Query by visual example, content based image retrieval," Advances in Database TechnologyEDBT'92, vol. 580, pp. 56-71, A. Pirotte, C.Delobel, and G. Gottlob, Eds., 2006, Springer-Verlag.
[4] J. Jeon, V. Lavrenko, and R. Manmatha, "Automatic Image Annotation and Retrieval Using Cross-Media Relevance
Models," Proc. 26th Ann. Int'l ACM SIGIR Conf. Research and Development in Information Retrieval (SIGIR '03),
pp. 119-126, 2003.
[5] J. Kalervo, K. Jaana, and N. Timo, "ExpansionTool: Concept-Based Query Expansion and Construction,"
Information Retrieval, vol. 4, no. 3, pp. 231-255, 2001.
[6] G. Chen, X. Li, L. Shou, J. Dong, and C. Chen, "HISA: A Query System Bridging the Semantic Gap for Large Image
Databases (Demo)," Proc. 32nd Int'l Conf. Very Large Data Bases (VLDB '06), pp. 1187-1190, 2006.
Volume 3, Issue 5, May 2014
Page 116
International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org
Volume 3, Issue 5, May 2014
ISSN 2319 - 4847
[7] X.Y. Li, L.D. Shou, G. Chen, and K.-L. Tan, "An Image-Semantic Ontological Framework for Large Image
Databases (Poster)," Proc. 12th Int'l Conf. Database Systems for Advanced Applications (DASFAA '07), pp. 10501053, 2007.
[8] J.R. Paek and S. Smith, "Detecting Image Purpose in World-Wide Web Documents," Proc. IS&T/SPIE Symp.
Electronic Imaging: Science and Technology-Document Recognition, vol. 3305, pp. 151- 158, Jan. 1998.
[9] A.W.M. Smeulders, M. Worring, S. Santini, A. Gupta, and R. Jain, "Content-Based Image Retrieval at the End of
the Early Years," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 22, no. 12, pp. 1349-1380, Dec. 2000.
AUTHOR
Sagar Ghan is persuing the B.E. degree in Information Technology from Sinhgad Institute of Technology
and Science (SITS).During 2013-2014, he studied various strategies of image processing and is undergoing
research in the field of image processing using CBIR technique.He along with other authors have developed
an efficient CBIR system.
Volume 3, Issue 5, May 2014
Page 117