Image searching based on image mean distance method

Image searching is most interesting in the field of the computer vision. Every day many digital images are coming into the web. User is attracted to automatic image retrieval from this large dataset. Many methods which are introduced in this last ten years for image retrieval based on the similarity, size of database, image classification, similar group of images finding, performance of retrieval. when the size of database is increasing image similarity finding It is big task for the researchers to give the efficient solutions. Content-Based Image Retrieval (CBIR) systems are used in order to retrieve image from image dataset. In our proposed method, we are utilizing clustering method for retrieve the images. Each Image converted into gray form and calculated the threshold values for 10000 images for each image extracted cluster mean values using k-mean method and stored into the database. In first method query image mean values are comparing using mean distances between query image and database images. In second method group of images are selecting from the image database with similar features and finding the distance difference of the each image and considered the squires of the all distances and make a sum and divide with the number of cluster values then make in sequential order of the images. Main aim of this work is to extract images with good similarity when the images that are retrieved based on query image. Above mention methods are given good performance.