A Dataset on Digital Image Forgery Detection

As per Nation Crime Record Bureau, 272 suicidal deaths have been recorded in India during 2021–22 due to the manipulation of images. Image has a remarkable role in various areas such as: Forensic investigation, Support to Police, LEA, and Judges to solve real cases, Teaching, Digital Image Forgery Detection. Digital image processing has become quite a simple operation for even an untrained user due to the accelerated technological advancements in digital image processing and consequently the rising use of digital recording equipments. Using modern digital image treatment software, every user can modify digital images in a way that makes it almost impossible to distinguish a fake from the first information visually. This paper presents a novel dataset for image manipulation detection. The increasing number of instances of tampered images being used as evidence in newspapers and courts has raised concerns about image forgery. ML was used to automatically generate semantically meaningful forgeries for the dataset. There have been four distinct types of forgeries produced. Forgeries known as splicing involve inserting an external element into an image, forgeries known as copy–move, in which an element within an image is duplicated, forgeries known as object removal, and morphing, in which two images are warped and blended together. As many as 5000 pictures have been created and each picture is joined by a few comments permitting exact limitations of the fabrication and data about the altering process. For the identification of copy–move, and splicing forgery recognition, certain helpful strategies are accessible. This paper presents a dataset on image forgery detection using ML. We will first introduce forensic digital technology, their formats and how they are compressed, subsequently highlight some important tools that were implemented to detect image forgery, and how well these tools can detect fake or tampered images.