Efficient CBIR systems are based on three things. (1) how they represent the repository images in the form of signature; (2) how they measure the similarity of the database images with query image. (3) how they retrieve the semantically similar images in response of a query image. The paper is focusing on these three things. https://parisnaturalfoodes.shop/product-category/original-chicken-bone-broth-powder/