Color Image Segmentation and Recognition based on Shape and Color Features Rajivkumar Mente, Department of Computer Science, Solapur University, Solapur rajivmente@rediffmail.com B. V. Dhandra, Gururaj Mukarambi, Department of Computer Science, Gulbarga University, Gulbarga Abstract – Recent advances in computer technology have made it possible to create databases for large number of images. A major approach directed towards achieving CBIR is the use of low-level visual features of the image data to segment, index and retrieve relevant images from the image database. Segmentation is the partition of a digital image into regions to simplify the image representation into something that is more meaningful and easier to analyze. Color based segmentation is significantly affected by the choice of color space. In different color spaces, the L*a*b color space is a better representation of the color content of an image. In this paper the L*a*b color space and K- means algorithm is used for segmentation of color images. Shape description or representation is an important issue both in object recognition and classification. After segmentation this paper focuses on the shape descriptor- eccentricity and color features for achieving efficient and effective retrieval performance. The proposed method is applied to an image database containing 2600 fruit images. Keywords – Segmentation, Clustering, L*a*b, Color space, Eccentricity, kNN, Canny Edge Detection I. INTRODUCTION “Segmentation” refers the process of dividing an image into distinct regions with property that each region is characterized unique feature such as intensity, color etc. Further it refers to the process of dividing a digital image into multiple segments such as sets of pixels [1]. Image segmentation may be defined as a process of assigning pixels to homogenous and disjoint regions which form a partition of the image that share certain visual characteristics [2]. It basically aims at dividing an image into subparts based on certain feature. Features could be based on certain boundaries, contour, color, intensity or texture pattern, geometric shape or any other pattern[3]. The result of image segmentation is a set of segments that collectively cover the entire image, or a set of contours extracted from the image[4]. Region growing approach used for segmentation examines neighboring pixels of initial “seed points” and determines whether the pixel neighbors should be added to the region[5]. Segmentation is application dependent because it needs image interpretation (e.g. locating objects or object boundary, lines etc). So it can be used in applications which involve a particular kind of object recognition such as  Multimedia applications  Industry applications (Airport security system, Automatic car assembly in robotic vision)  Face recognition  Fingerprint recognition  Locating objects in satellite images In the color image segmentation, a proper choice of color space is very important. For experiment CIE L*a*b color space is chosen due to its three major properties 1. Separation of achromatic information from chromatic information. 2. Uniform color space and 3. Similar to human visual perception. Here L* represents the luminance component, while a* and b* represent color components[6]. The most popular method for image segmentation is K-means Clustering algorithm. It is a widely used algorithm for image segmentation because of its ability to cluster huge data points very quickly. [7][8][9]. Applications like medicine, entertainment, education etc. make of vast amount of visual data in the form images. The features such as luminance, shape descriptor and gray scale texture are some natural features since Rajivkumar Mente et al. / International Journal of Computer Science Engineering (IJCSE) ISSN : 2319-7323 Vol. 3 No.01 Jan 2014 51