Content Based Image Retrieval System using HSV Color Space to Support Fake News Analysis

The Content Based Image Retrieval (CBIR) system is applicable for the retrieval of digital images by using their visual contents such as color, shape and texture. In our proposed we had used color as a visual content for the retrieval of images. HSV color space has been used to extract the color feature from the image. This paper present an approach to overcome from the semantic gap challenge which comes from the high-level human semantics. This paper present a technique to overcome from such challenge and the technique is to applied an elliptical mask on the image and divide the image into five regions i. e, top left, top right, bottom left, bottom right and centre of the image. For the similarity matching chi-square distance has used. This paper also present an application implementation of the CBIR system for the analysis of fake news. The retrieval results were obtained by applying the HSV histogram on the dataset of around 1025 images of various categories like flower, buses, politician, Egyptian pyramid and nature related images.