INTEGRATED APPROACH TO CONTENT BASED IMAGE RETRIVAL USING CLUSTERING

Abstract

Content-based image retrieval has, over the past few years, got a lot of thought. Content-Based Image Retrieval (CBIR) is essentially a strategy to perform image retrieval from a huge database similar to images delivered on demand. CBIR is closer to human semantics, as it relates to measuring image retrieval. The CBIR system is applied in various places, for example, bad behavior prevention, clinical imagery, weather prediction, recognition, chronic investigation, and remote recognition. The content here refers to the visual data of the images, for example texture, shape, and color. Image content is more data extreme to interestingly recover capability with text-based image recovery. In this article, we suggested a combined content-based photo recovery technique that isolates both the color and texture feature. To remove the color feature, the color moment (CM) is used in the color images and to isolate the texture feature, the local binary pattern (LBP) is performed on the grayscale image. By then, the image color and texture feature are fused together to form a single feature vector. Ultimately, similarity is regulated by Euclidean distance which compares the image properties vector in the database with the question images. LBP is primarily used for facial recognition. Anyway, we'll use LBP for regular photos. This combined strategy provides a precise, beneficial, and less complex recovery framework.