# Introduction ontent Based Image Retrieval is a term used to describe the process of retrieving images form a large collection on the basis of features(such as COLOUR, texture etc) that can be automatically extracted form the images themselves [19]. The retrieval thus depends on the contents of images. CBIR is comparatively a new topic and has turn into a hot research subject in last few years. The features that are used in CBIR should correspond directly to general routine notions of the human vision. For example, COLOUR, texture and shape are general terms used by most of people. These features are mostly used for CBIR. But it is hard to define precisely how these features are discriminated by humans. Also, such discriminations are different for different people. Thus we need to pre-define the suitable feature representation scheme for each of these features. These features can be applicable over complete image or over a small region of the image. It is found that in most image retrieval systems, COLOUR based features play a prominent role. Indeed, COLOUR is the most important factor in human perception. The majority common representation of COLOUR information is in the form of COLOUR histogram, which statically is the probability of any given pixel having a Specific intensity in each of the COLOUR channels. COLOUR anglogram [20], correlogram [18], COLOUR co occurrence matrix (CCM) [22] are some of the other feature representations for COLOUR. # II. # System Model In content-based image retrieval (CBIR) image databases are indexed with descriptors derived from the visual content of the images. Most CBIR systems are worried with estimated queries where the aim is to find pictures visually similar to a particular target picture. In most cases the goal of CBIR systems is to replicate human perception of image similarity as well as possible. CBIR presents a challenging problem since it has common elements with both the general image understanding problem (which seems to remain unsolvable for computers at least in the near future) and the field of general information retrieval. Humans excel in image understanding when compared with computers. In contrast, in systematic handling of large databases computers have an edge over us. For this reason CBIR is also potentially very rewarding. Global histogram gives information about the COLOUR contents of image and not the spatial distribution of COLOUR in image. Thus COLOUR feature alone cannot give satisfactory results. Texture can give additional information about the spatial arrangements and patterns of varying intensity available image. Texture is an important element to human vision. Texture has been found to provide cause to scene depth and surface orientation. People also tend to relate texture elements of varying size to a 3-D surface. Even in graphic systems greater realism is achieved when textures are mapped to 3-D surfaces. Gabor filters, Tamura filters, Gray level co occurrence matrix (GLCM) etc are used for the texture representation. Discrete Wavelet Transform (DWT) is found to be an effective tool for signal analysis. Wavelets have properties that are suitable for representation of an image texture. Daubechies wavelets and Haar wavelets are some of the examples of texture representation [19]. Shape information is considered to be one of the most difficult C features to extract reliably from images since there are no mathematical definitions of shape similarity which Year 2014 F assign to shape. While using shape, it is important that the representation should be invariant to basic transformations such as rotation, scale etc. some of the features that are used to represent shape are moment invariants, circularity, area, minimum and maximum axis. These features of the database images are calculated offline and stored into database. User gives one image as the query image (which may be uploaded or selected from the previous results). Then a system calculates corresponding features for the query image. Similarity measure takes these feature values and those from database and calculate the similarity among query image and each of the database images. Database images finally are ranked and displayed according to similarity measure. In general, CBIR can be described in terms of following stages: # Proposed Methodology The image retrieval system proposed in this paper is described below with the block diagram and flow charts. In Fig. 3.1 block diagram of feature database preparation is described. in Fig. 3.2 block diagram of image retrieval is proposed. In Fig. 3.3 flow chart of feature database preparation algorithm is explained. In Fig. 3.4 flow chart of image retrieval algorithm is explained. In the block diagrams the proposed methodology is described in first figure the Feature Database is prepared by using the collection of images, features extraction and pattern with color so the preparation of database has been done and then after store in Database. In this figure the Image Retrieval process is described by using the Feature, Extract the Features from image and compare with Feature Data Base then prepare the similarity table at the end 16 In the first flow graph preparing the features database process is achieved in which all images is loaded in PC and also feature database is loaded then resizing to extract feature is done then after Median Filter is adopted for removing the noise then extraction of colour feature from images is done then find out the dominant colour percentage and maximum value is achieved then extraction of pattern feature and store into database. # Figure 3.4 : Flow char of image retrieval b) Image Retrievals In this flow graph the image retrieval process is achieved. First of all browsing the Query Image then extract feature (Pattern with Color) of Query Image then Loading of Features Database for Comparison purpose then Comparing Features of Query Image after that list is prepared for Feature Similarity then short features in descending order and finally the results have been displayed. IV. # Simulation Results Image retrieval proposed methodology is implemented on simulation tool and various results are found. The main reason is to develop advanced technique is to get the images from the huge collection faster than every other technique with meaningful results. In this section simulation results of proposed method is shown with different color and shapes. In Fig. 4.1 proposed algorithm trying to retrieve images of red colour looking similar to girl and the similar results are displayed in the figure. In the retrieved results algorithm first trying to find out the similar images in descending order that is most similar images appear first in the results. The time taken to get the results is 1.8268 seconds. In Fig. 4.2 proposed trying to retrieve images of green colour looking similar to cup and the similar results are displayed in the figure. In the retrieved results algorithm first trying to find out the similar images from images collection and arrange in descending order i.e. most similar images appear first in the results. The time taken to get the results is 1.8207 seconds. In the retrieved results algorithm first trying to find out the similar images from images collection and arrange in descending order i.e. most similar images appear first in the results. The time taken to get the results is 1.8406 seconds. As we were tried for other shapes and colours the results algorithm takes on an average 1.8 seconds to retrieve each images for each query. V. # Conclusion and Future Scope Image retrieval is growing and in demand technique used in wide area of application like search engines, social networking sites, surveillance systems etc. The need of image searching is either colour based or patterns. In the proposed methodology of this paper we have adopted method for colour based retrieval as well as pattern based retrieval and we have tried to make is faster like 1.8 seconds to get results. In the upcoming time hybrid form of multiple techniques definitely improve the accuracy as well as retrieval time. ![Author ? ? ?: M-Tech Research Scholar, Research Guide, Department of Computer Science Engineering Technocrats Institute of Technology (excellence), Bhopal. e-mails: gayatritiwari.be2012@gmail.com, bhupesh_gour@rediffmail.com, asifullahkhan@rediffmail.com](image-2.png "&") ![a) Identification and utilization of intuitive visual features. b) Features representation c) Automatic extraction of features. d) Efficient indexing over these features. e) Online extraction of these features from query image. f) Distance measure calculation to rank images. III.](image-3.png "") 31![Figure 3.1 : Block Diagram of Feature Database is preparation](image-4.png "Figure 3 . 1 :") 32![Figure 3.2 : Block Diagram of Image Retrieval](image-5.png "Figure 3 . 2 :") CollectionofImagesFeatures ExtractionPrepare Feature DatabaseStore DBFeatures DatabasePattern withColorLoadingFeatureFeatureDataInput Query ImageExtract FeaturesCompare Features with DBPrepare Similarity TableSort SimilarityOutput Top16 Results © 2014 Global Journals Inc. (US) * Image Database Systems: A Survey HTamura NYokoya Pattern Recognition 17 1 Sep. 1984 * Term-Weighting Approaches in Automatic Text Retrieval SGerard CBuckely Information Processing and Management 24 5 Jan. 1988 * Image Categorization by Learning and Reasoning with Regions YChen JWang Journal of Machine Learning Research 5 May 2004 * Fundamentals of content based image retrieval FLong HZhang HDagan DFeng Multimedia Information Retrieval and Management D. Feng, W. Siu, H. Zhang * Technological Fundamentals and Applications 2003 Springer-Verlag Berlin Heidelberg New York Multimedia Signal Processing Book * Content-based image retrieval systems VGudivada Raghavan IEEE Computer 28 Sep. 1995 * Image Retrieval From the World Wide Web: Issues, Techniques, and Systems MKherfi DZiou ABernardi ACM Computing Surveys 36 1 March 2004 * Query by image and video content: The QBIC system MFlickner HSawhney WNiblack JAshley QHuang BDom MGorkani JHafner DLee DPetkovic PYanker IEEE Computer 28 Sep. 1995 * Photobook: Content based manipulation of image databases APentland RPicard SSclaroff International Journal of Computer Vision 18 June 1997 * Visualseek: A Fully Automated Content-Based Image Query System JSmith SChang Proceedings of the 4th ACM international conference on Multimedia table of contents the 4th ACM international conference on Multimedia table of contentsBoston, Massachusetts, United States Nov. 1996 * Visual information retrieval AGupta RJain Comm. Assoc. Comp. Mach 40 5 May. 1997 * Integrated Region Matching for Image Retrieval JLi JWang GWiederhold Proceedings of the 2000 ACM Multimedia Conference the 2000 ACM Multimedia ConferenceLos Angeles October 2000 * A new feature set for content based image retrieval MBRao CKavitha BPRao AGovardhan 2013 International Conference on Feb. 2013 89 Information Communication and Embedded Systems (ICICES * LunshaoChai Zhen Qin * HonggangZhang Jun Guo * Re-ranking using compressionbased distance measure for Content-based Commercial Product Image Retrieval CRShelton 19th IEEE International Conference on 2012. 1944. Sept. 30 2012-Oct. 3 2012 1941 * HaipengLi * JianboLu * An image retrieval method based on fractal image coding JiangYu Systems and Informatics (ICSAI), 2012 International Conference on 1869, 19-20 May 2012 1866 * Flickr Retriever --Fast Retrieval of Flickr Photos DEdmundson GSchaefer Web Intelligence and Intelligent Agent Technology (WI-IAT) 2012 * Ieee/Wic/ ACM International Conferences on 1 Dec. 2012 * Image retrieval based on content and image compression ASAbdelrahim MAAbdelrahman AMahmoud AAFarag Multimedia Technology (ICMT), 2011 International Conference on July 2011 6703 * The application of fractal compression to content-based image retrieval: Comparison of techniques An-ZenShih Multimedia Technology (ICMT), 2010International Conference on 26 Apr 2010 6484 3356 * Image Indexing using Color Correlograms JHuang SRKumar MMitra WJZhu RZabih IEEE Computer Society Conference on Computer Vision and Pattern Recognition June 1997 * A Wavelet Based Technique for Analysis and Classification of Texture Images SSiddique April 2002 Ottawa, Canada Carleton University M.S Thesis * Narrowing the Semantic Gap-Improved Text Based Web Document Retrieval Using Visual Feature RZhao WIGrosky IEEE Transactions on Multimedia 4 2 June 2002