Content Based Image Retrieval Using SVM

Content based image retrieval has been emerged as the most innovative area for the research work from the last few years. Image retrieval is basically a method for the extract ion of most crucial characteristics of an image. The primary aim of th is content based image retrieval (CBIR) is to have the exact result within the short period of t ime. Content-based image retrieval is based on the concept of using the depiction of characteristics which are retrieved from the images themselves. Maximum of the present CBIR systems works on querying-by-example, an approach in which a portion of the image or the whole image is chosen by the user in the form of query. The technique works by retrieving the feature of the image which is given as a query then going through the database for the image having similar features as the image in the query and ext racts the appropriate images to the user which matches most appropriately to the query. In this area, content in itself constitutes the other featu res also like the texture of the image, co lor of the image, kinesthetic relationships and shape of the image. A number of CBIR systems have been studied earlier which makes the use of these features for the purpose of comparison, analyzing and extract ion of image. There are also some systems that have also been successful by integrating the both content -based retrieval as well as the text-based retrieval technique. In all situations, however, it has been not possible to conclude that which particu lar feature is evolved as the most important for impart ing the best extract ion. Many researchers used various technique s for the classificat ion of huge amount of data. In this paper we emphasize on a SVM (Support Vector Machine) technique for the purpose of retrieval of images similar to the image in the query. In this paper we primarily present the importance or necessit y of support vector machine. The tradit ional content –based image retrieval techniques may not work as required because of some problems in real world application. For example all users would like to get their desired result in the first time on the web, p ossibility of consuming a lot of time while searching for the results. In this paper we contemplate over the mechanism of support vector machine (SVM) and also explain some other methods for extracting images.