Main Article Content
An important facet of traditional retrieval models is that they retrieve images and videos and consider their content and context reliable. Nevertheless, this consideration is no longer valid since they can be faked for many reasons and at different degrees thanks to powerful multimedia manipulation software. Our goal is to investigate new ways detecting possible fake in social network platforms. In this paper, we propose an approach that assets identification faked images by combining standard content-based image retrieval (CBIR) techniques and watermarking. We have prepared the wartermarked image database of all images using LSB based watermarking. Using gabor features and trained KNN, user is able to retrieve the matching query image. The retrieved image is authenticated by extracting the watermark and matching it again with the test image.