1.

INTRODUCTION
1.1 Definition of image
Data representing a two-dimensional scene. A digital image is composed of pixels arranged in a rectangular array with a certain height and width. Each pixel may consist of one or more bits of information, representing the brightness of the image at that point and possibly including color information encoded as RGB triples. Picture a visual representation (of an object or scene or person or abstraction) produced on a surface, "they showed us the pictures of their wedding" a movie is a series of images projected so rapidly that the eye integrates them. Effective indexing and retrieving desired image in large image database in the basis of features such as color, text and shape that can be automatically extracted from the images themselves.

1.2 Image processing

The analysis of a picture using techniques that can identify shades, colors and relationships that cannot be perceived by the human eye. Image processing is used to solve identification problems, such as in forensic medicine or in creating weather maps from satellite pictures. It deals with images in bitmapped graphics format that have been scanned in or captured with digital cameras. Any image improvement, such as refining a picture in a paint program that has been scanned or entered from a video source.

1|Page

1.3 Image Retrieval and Information Retrieval

Since the 1970s Image Retrieval has become a very active research topic, with two major research communities, database management and computer vision. One is text-based and another is visual-based. Text-based image retrieval has become very popular since 1970s, which involves annotating the image with keywords, and use text-based database management systems (DBMS) to retrieve the images. In text-based image retrieval system, keywords of semantic information are attached to the images.

They can be typed manually or by extracting the captions of the images. It is very efficient for simple and small image databases, since the whole database can be described by just few hundreds of keywords. But in the 1990s, several large leaps in development of processor, memory and storage made the size of image databases grow dramatically. As the image database and image size grow, there will be more images having different contents and the images having rich contents cannot be described by only several Semantic keywords. The demand of labor on annotating the images also rises dramatically.

Retrieval image providing effective and efficient tool querying large image database. Information retrieval provides the textual representation of images. It requires the text descriptions to the respective images.

Recent technology development in various fields has made large digital image databases practical. Well organized database and efficient browsing, storing, and retrieval algorithms are very important in such systems. Image retrieval techniques were developed to aid these components.

2|Page

Image Retrieval was originated from Information Retrieval, which has been very active research topic since 1940s. We have huge amounts of information to which accurate and speedy access is becoming ever more difficult. In principle, Information Retrieval is simple. It can be illustrated by a scene of a store of documents and a person (user of the store). He formulates a question to which the answer is a set of documents satisfying his question. He can obtain the set by reading all the documents in the store, retaining the relevant documents and discarding all the others. In this scene, it is a perfect retrieval. But in practice, we need to model the read process in both syntactic and semantic to extract useful information. The target of Information Retrieval is not only how to extract useful information, but also how to measure relevance among documents. These challenges also exist in Image Retrieval. Also the keywords are very dependent on the observers interest and they are subjective. Captions are not always precisely describing the picture. Indexing and searching a large image database via keywords are time-consuming and inefficient. Content Based Image Retrieval (CBIR) researches attempt to automate such complex process of retrieving images that are similar to the reference image or descriptions given.

1.4 Content Based Image Retrieval

Content-based image retrieval also known as query by image content and content-based visual information retrieval problem of searching for digital images in large database. Content-based means that the search will analyze the actual contents of image. The term content in this context might refer to colors, shapes, textures or any other information that can be derived from the image itself.
3|Page

The earliest use of the term Content Based Image Retrieval in the literature seems to be by Kato, was to describe his experiments in automatic retrieval of images from a database by color and shape features. The term has since been widely used to describe the process of retrieving desired images from a large collection on the basis of features (such as color,Texture and shape) that can be automatically extracted from the images themselves. The features used for retrieval can be either primitive or semantic, but the extraction process must be predominantly automatic.

The ideal approach of querying an image database is using content semantics, which applies the human understanding about image. Unfortunately, extracting the semantic information in an image efficiently and accurately is still a question. Even with the most advanced implementation of computer vision, it is still not easy to identify an image of horses on a road. So, using low level features instead of semantics is still a more practical way. Until semantic extraction can be done automatically and accurately, image retrieval systems cannot be expected to find all correct images. They should select the most similar images to let the user choose the desired images. The number of images of retrieved set can be reduced by applying similarity measure that measures the perceptual similarity.

A typical CBIR system consists of three major components and the variations of them depend on features used. i. Feature extraction Analyze raw image data to extract feature specific Information. ii. Feature storage Provide efficient storage for the extracted information, also help to improve searching speed.

4|Page

iii. Similarity measure Measure the difference between images for determining the relevance between images.

1.5 Content Based Image Retrieval using color

Retrieving image based on color similarity is achieved by computing a color histogram for each image that identifies the proportion of pixels within an images holding specific values. Current research is attempting to segment color proportion by region and by spatial relationship among several color region.

Among different types of low level features, color is the most straightforward information which can be easily retrieved from digital images with simple and compact description, while others require more pre-processing and computational tasks such as pattern recognition or texture analysis.

While comparing image by color feature, three properties are usually considered: i. Area of matching Count the area or number of pixels having same or similar colors. Larger matched area means more similar. ii. Color distance Distance between colors, usually in a perceptually uniform color space. Closer between matched colors means more similar. iii. Spatial distribution Usually used while combining color with other features such as texture and shape.

In a typical color similarity measure, area of matching is usually counted as the similarity color distance is used to control the matching between colors and to
5|Page

adjust the similarity. In conventional color image retrieval system, the most straight forward approach is using color histogram. Histograms of each color, for example, images of 256 colors, will be generated. Similarities between such images are then performed and measured by Histogram Intersection Method (HIM). This is the basic approach and can give simple and efficient representation of color distribution. Histogram approach is not limited by taking the number of pixels of each color in the image or using HIM similarity measure. Indexes of histograms can represent many types of features such as colors in different color space, coefficients in transformed domain or spatial-related information. There are also many variations in comparing histograms. But histograms have a limitation that the feature space is fixed, compactness of the description is limited, because histograms will not skip non-existed colors. One argument that we can use a lower resolution histogram to improve the compactness but it is a trade off between compactness and accuracy.

1.6 JPEG visual descriptors

That shape often carries semantic information follows from the fact that many characteristic objects can be visually recognized solely from their shapes. This distinguishes shape from other elementary visual features such as color, or texture. But the notion of object shape has many meanings. To deal with 3D realworld objects, JPEG standard has a 3D shape descriptor.

JPEG has defined a set of standard descriptors for description and storage of the most commonly used features. This makes the extracted features more accessible. Since the required storage size is much smaller than compressed images files. Moreover, the format of the data is fixed, so the data can be used in any
6|Page

JPEG compatible systems. Thus comparison between algorithms can be done easily if the implementations of the target Algorithms are JPEG compatible.

In JPEG visual standard, some color descriptors are defined, including several histogram based descriptors representing different color features, and a Dominant Color Descriptor (DCD). DCD describes color feature by a set of representative colors with their percentage and each color have at least a certain distance away in CIE color space controlled by a threshold Td. It is very compact since there is no redundant information for non-existed colors, and similar colors are grouped into a palette color.

1.7 Relevance Feedback

Although JPEG defined efficient and most commonly used CBIR methods, content based methods still have limitations that they may not be able to find the images that exactly match users expectation. One reason is that a precise query cannot be formulated Although DCD can describe color features in a compact and effective way, Drawbacks of its default similarity measure method pull down the performance of DCD.

By just giving an image as query, Interactive searching may be used for improving the retrieval result by refining the query by users feedback. JPEG did not handle interactive searching directly. They use content management approach to describe multimedia contents in a structural format. This uses textual semantics. This can improve the efficiency of browsing and text based searching, but not for content based searching. Also it is not interactive and the improvement may not show instantly. For content based interactive searching, Relevance Feedback (RF)
7|Page

is a commonly used technique which use user selected relevance information to refine the query.

It can be used as an extension of similarity measure and will not affect JPEG standard. Since the representations depend on the features, relevance feedback algorithms may be different for each descriptor. In this research, a merged palette histogram approach is proposed to improve Dominant Color Descriptor searching with use of RF.

But the effect of this coefficient is still not very clear. It is unable to balance the effect of area of matching and distance between colors.

It causes ambiguous retrieval results. These problems will be described in chapter 4, and a new Merged Palette Histogram Similarity Measure (MPHSM) will be proposed to tackle these problems.

8|Page

2. LITERATURE SURVEY

2.1 Introduction
In earlier days, image retrieving from large image database can be done by following ways. We will discuss briefly about the image retrieving of various steps : Automatic Image Annotation and Retrieval using Cross Media Relevance Models Concept Based Query Expansion Query System Bridging The Semantic Gap For Large Image Databases Ontology-Based Query Expansion Widget for information Retrieval Detecting image purpose in World-Wide Web documents

2.2 Automatic Image Annotation and Retrieval using Cross Media Relevance Models
Libraries have traditionally used manual image annotation for indexing and then later retrieving their image collections. However, manual image annotation is an expensive and labor intensive procedure and hence there has been great interest in coming up with automatic ways to retrieve images based on content [5]. Here, we propose an automatic approach to annotating and retrieving images based on a training set of images.

We assume that regions in an image can be described using a small vocabulary of blobs. Blobs are generated from image features using clustering.
9|Page

Given a training set of images with annotations, we show that probabilistic models allow us to predict the probability of generating a word given the blobs in an image. This may be used to automatically annotate and retrieve images given a word as a query. We show that relevance models. Allow us to derive these probabilities in a natural way. Experiments show that the annotation performance of this cross-media relevance model is almost six times as good (in terms of mean precision) than a model based on word-blob co-occurrence model and twice as good as a state of the art model derived from machine translation. Our approach shows the usefulness of using formal information retrieval models for the task of image annotation and retrieval.

2.3 Concept Based Query Expansion

Query expansion methods have been studied for a long time - with debatable success in many instances. In this project we present a probabilistic query expansion model based on a similarity thesaurus which was constructed automatically[6]. A similarity thesaurus reflects domain knowledge about the particular collection from which it is constructed. We address the two important issues with query expansion: the selection and the weighting of additional search terms. In contrast to earlier methods, our queries are expanded by adding those terms that are most similar to the concept of the query, rather than selecting terms that are similar to the query terms.

Our experiments show that this kind of query expansion results in a notable improvement in the retrieval effectiveness when measured using both recallprecision and usefulness.

10 | P a g e

2.4 Query System Bridging the Semantic Gap for Large Image Databases
We propose a novel system called HISA for organizing very large image databases. HISA implements the first known data structure to capture both the ontological knowledge and visual features for effective and efficient retrieval of images by either keywords, image examples, or both. HISA employs automatic image annotation technique, ontology analysis and statistical analysis of domain knowledge to precompile the data structure [7]. Using these techniques, HISA is able to bridge the gap between the image semantics and the visual features, therefore providing more user-friendly and high performance queries. We demonstrate the novel data structure employed by HISA, the query algorithms, and the pre-computation process.

2.5 Ontology-Based Query Expansion Widget for information Retrieval

In this project we present an ontology-based query expansion widget which utilizes the ontologies published in the ONKI Ontology Service. The widget can be integrated into a web page, e.g. a search system of a museum catalogue, enhancing the page by providing query expansion functionality. We have tested the system with general, domainspecic and spatiotemporal ontologies.

11 | P a g e

2.6 Detecting image purpose in World-Wide Web documents

The number of World-Wide Web (WWW) documents available to users of the Internet is growing at an incredible rate. Therefore, it is becoming increasingly important to develop systems that aid users in searching, altering, and retrieving information from the Internet. Currently, only a few prototype systems catalog and index images in Web documents. To greatly improve the cataloging and indexing of images on the Web, we have developed a prototype rule-based system that detects the content images in Web documents [9]. Content images are images that are associated with the main content of Web documents, as opposed to a multitude of other images that exist in Web documents for different purposes, such as decorative, advertisement and logo images. We present a system that uses decision tree learning for automated rule induction for the content image detection system. The system uses visual features, text-related features and the document context of images in concert for fast and effective content image detection in Web documents.

2.7 Content Based Image Retrieval

Content Based Image Retrieval (CBIR) [10] 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
12 | P a g e

representations, histogram is the most commonly used technique to describe features.

Fig 2.1 Flow of a typical CBIR process

Fig 2.1 describes the flow of a typical CBIR process although content based methods are efficient, they cannot always match users expectation. Relevance Feedback (RF) techniques are used to adjust the query by users feedback. RF is an interactive process to improve the retrieval accuracy by a few iterations. RF algorithms are dependent on feature representations, in this chapter, RF process and its histogram weighting method will be introduced.

13 | P a g e

3. PROBLEM STATEMENT
It aims to introduce the problems and challenges concerned with the design and the creation of CBIR(Content Based Image Retrieval) systems, which is based on a free hand sketch (Sketch based image retrieval SBIR).

14 | P a g e

4. HARDWARE AND SOFTWARE REQUIREMENTS

4.1 Hardware Requirements
System Hard Disk Floppy Drive Monitor RAM

: Pentium IV 2.4 GHz. : 40 GB. : 1.44 Mb. : 15 VGA Colour. : 256 Mb.

4.2 Software Requirements

Operating system Front End Tool

:- Windows XP Professional :- JAVA, Swing :- Eclipse 5.0

15 | P a g e

6. IMPLEMENTATION
Implementation is the stage of the project when the theoretical design is turned out into a working system. Thus it can be considered to be the most critical stage in achieving a successful new system and in giving the user, confidence that the new system will work and be effective. The implementation stage involves careful planning, investigation of the existing system and its constraints on implementation, designing of methods to achieve changeover and evaluation of changeover methods.

4.5 1 Introduction
Indexing the whole set of images using K-means Clustering algorithm. Indexing is done using an implementation of the Document Builder Interface. A simple approach is to use the Document Builder Factory, which creates Document Builder instances for all available features as well as popular combinations of features (e.g. all JPEG features or all avail-able features).

In a content based image retrieval system, target images are sorted by feature similarities with respect to the query (CBIR).In this indexing, we propose to use Kmeans clustering for the classification of feature set obtained from the histogram. Histogram provides a set of features for proposed for Content Based Image Retrieval (CBIR). Hence histogram method further refines the histogram by splitting the pixels in a given bucket into several classes. Here we compute the similarity for 8 bins and similarity for 16 bins. Standard histograms, because of their efficiency and insensitivity to small changes, are widely used for content
16 | P a g e

based image retrieval. But the main disadvantage of histograms is that many images of different appearances can have similar histograms because histograms provide coarse characterization of an image.

The k-means algorithm takes the input parameter, k, and partitions a set of n objects into k clusters so that the resulting intra cluster similarity is high but the inter cluster similarity is low. Cluster similarity is measured in regard to the mean value of the objects in a cluster, which can be viewed as the clusters center of gravity. How does the k-means algorithm work.

The k-means algorithm proceeds as follows. First, it randomly selects k of the objects, each of which initially represents a cluster mean or center. For each of the remaining objects, an object is assigned to the cluster to which it is the most similar, based on the distance between the object and the cluster mean. It then computes the new mean for each cluster.

The k-means algorithm:

Algorithm: k-means. The k-means algorithm for partitioning based on the mean value of the objects in the cluster.

Input: The number of clusters k and a database containing n objects. Output: A set of k clusters that minimizes the squared-error criterion. Method: (1) arbitrarily choose k objects as the initial cluster centers: (2) repeat

17 | P a g e

(3) (re)assign each object to the cluster to which the object is the most similar, based on the mean value of the objects in the cluster; (4) Update the cluster means, i.e., calculate the mean value of the objects for each cluster; (5) Until no change.

Efficiency :1) Parallel Implementations 2) Reduction of distance computations 3) Scaling strategies

Complexity :- The complexity of the K-means Algorithms is O( n * K * I * d ) where n = number of points, K = number of clusters, I = number of iterations, d = number of attributes

18 | P a g e

CONCLUSION
Among the objectives of this paper performed to design, implement and test a sketch-based image retrieval system. Two main aspects were taken into account. The retrieval process has to be unconventional and highly interactive. The robustness of the method is essential in some degree of noise, which might also be in case of simple images. The drawn image without modification cannot be compared with color image, or its edge representation. Alternatively a distance transform step was introduced. The simple smoothing and edge detection based method was improved, which had a similar importance as the previous step. At the tests the effectiveness of EHD and the dynamically parameterized HOG implementation was compared. It was examined with more databases. In our experience the HOG in more cases was much better than the EHD based retrieval. However, the situation is not so simple. The edge histogram descriptor can mainly look better for information poor sketches, while in other case better results can be achieved for more detailed. This is due to the sliding window solution of HOG. Using the SIFT-based multi-level solution the search result list is renew. With the
19 | P a g e

categorization of retrieval response a bigger decision possibility was given to the user on that way, he can choose from more groups of results.

10. BIBLIOGRAPHY
Good Teachers are worth more than thousand books, we have them in Our Department [1] D. Comaniciu, and P. Meer, Robust analysis of feature spaces: color image segmentation, IEEE Conference on Computer Vision and Pattern Recognition, pp. 750755, June 1997. [2] N. Dalal, and B. Triggs, Histograms of oriented gradients for human detection, IEEE Conference on Computer Vision and Pattern Recognition, pp. 886893, July 2005. [3] T. Deselaers, D. Keysers, and H. Ney, Features for image retrieval: an experimental comparison, Information Retrieval, vol. 11, pp. 77107, December 2007.
20 | P a g e

[4] M. Eitz, K. Hildebrand, T. Boubekeur, and M. Alexa, An evaluation of descriptors for large-scale image retrieval from sketched feature lines, Computers and Graphics, vol. 34, pp. 482498, October 2010. [5] R. Fabbri, L.D.F. Costa, J.C. Torelli, and O.M. Bruno, 2D Euclidean distance transform algorithms: a comparative survey, ACM Computing Surveys, vol. 44, pp. 144, February 2008.

[6] M. Flickner, H. Sawhney, W. Niblack, J. Ashley, Q. Hiang, B. Dom, M. Gorkani, J. Hafner, D. Lee, D. Petkovic, D. Steele, and P. Yanker, Query by image and video content: the QBIC system, IEEE Computer, vol. 28, pp. 23 32, 2002. [7] Gy. Gyorok, Embedded hybrid controller with programmable analog circuit, IEEE 14th International Conference on Intelligent Systems pp. 59.1 59.4, May 2010. [8] R. Hu, M. Barnard, and J. Collomosse, Gradient _eld descriptor for sketch based image retrieval and localization, International Conference on Image Processing, pp. 14, 2010. [9] A.K. Jain, J.E. Lee, and R. Jin, Sketch to photo matching: a feature-based approach, Proc. SPIE, Biometric Technology for Human Identi_cation VII, vol. 7667, pp. 766702766702, 2010.

21 | P a g e

[10] A.K. Jain, J.E. Lee, R. Jin, and N. Gregg, Graf_ti-ID: matching retrieval of graf_ti images, ACM MM, MiFor09, pp. 16, 2009. [11] A.K. Jain, J.E. Lee, R. Jin, and N. Gregg, Content based image retrieval: an application to tattoo images, IEEE International Conference on Image Processing, pp. 27452748, November 2009

[12] T. Hashimoto, A. Rovid, G. Ohashi, Y. Ogura, H. Nakahara, and A.R. Varkonyi-Koczy, Edge detection based image retrieval method by sketches, Proc. of the Internatonal Symposium on Flexible Automation, pp. 1 4, 2006. [13] J.B. Kruskal, Nonmetric multidimensional scaling: a numerical method, Psychometrika, vol. 29, pp. 115129, 1964. [14] Y. Liu, and F. Dellaert, A classi_cation based similarity metric for 3D image retrieval, IEEE Conference on Computer Vision and Pattern Recognition, pp. 800805, June 1998. [15] D.G. Lowe, Distinctive image features from scale-invariant keypoints, International Journal of Computer Vision, vol. 60, pp. 91110, 2004. [16] D.G. Lowe, Object Recognition from Local Scale-Invariant Features, IEEE International Conference on Computer Vision, vol. 2, p. 1150, 1999.
22 | P a g e

[17] J.R. Smith, and S.F. Chang, VisualSEEK: a fully automated contentbased image query system, ACM Multimedia 96, pp. 9798, 1996. [18] J. Tick, and J. Fodor, Some classes of binary operations in approximate reasoning, Studies in Informatics and Control, vol. 15, pp. 259270, 2006. [19] S.J. Wan, P. Prusinkiewicz, and S.K.M. Wong, Variance-based color image quantization for frame buffer display, Color Research and Application, vol. 15, pp. 5258, February 1990.

Sites Referred:
http://java.sun.com http://www.sourcefordgde.com http://www.networkcomputing.com/ http://www.roseindia.com/ http://www.java2s.com/ http://www. javadb.com/

23 | P a g e