Detecting Deceptive Images in Online Content

Due to the availability and ease of advanced image editing techniques, individuals now can edit and alter images, sometimes in a way make it hard for the human eye to differentiate the altered images from the authentic ones. Moreover, recently the freedom and the simplicity of social network platforms facilities sharing and spreading online content with numerous suspicious images. As images have a powerful influence on a watcher's attitude and action towards the surrounded issue, therefore, there is a need to guarantee the trustworthiness of images attached to the online content. This research study aims to develop a system to automate authenticity analysis of images which helps users to detect altered images in online content. The developed system makes the use of both Error Level Analysis (ELA) and Clone detection algorithms to detect the modified parts in deceptive images and provide more clear explanation of the results. Steps has been added to the Error Level Analysis (ELA) algorithm to enhance the generated detection results. Based on the results from the developed system, we can conclude that our system can detect different image forgeries and display clear image analysis results for both deceptive and authentic images and had received a high level of satisfaction and willingness from different type of users.