Effective RBIR Fuzzy C-Means Segmentation Haar Wavelet with User Interactive Multi Threshold Robust Features Vector

摘要

Region-Based Image Retrieval (RBIR) is proposed as a development of Content-Based Image Retrieval (CBIR). In this study, an RBIR system Fuzzy C-Means (FCM) algorithm is proposed for image segmentation, it automatically segments images into a variable number of regions and extracts each region as a set of features. In this study, a user interactive multi threshold scheme for Region based image retrieval is proposed. We use speeded up robust feature technique to extract the flat level features of the image. We excerpt the robust features of the image, each contains 124 values which helps to reduce the dimensionality. For each image there will be number of feature vectors, the extracted features will be indexed for further retrieval of images. At the query phase, the same set of features will be extracted from the query image and will be compared with the indexed feature vectors, here we accepted methodologies for facilitating semantic image analysis process. In addition, the system provide a feed back information that is used to elaborate the retrieval results and the overall procedure emerged until the user regards the returned results as satisfactory. This algorithm improves the performance of image segmentation and the time efficient final results will be returned according to the threshold specified. The user will be allowed to provide threshold in order to get the new set of results.