Tampering Detection System ( A Machine Learning Approach )

With a big section of the population having access to the internet and a camera, the quantity of photographs shared online is growing exponentially. Using cutting-edge software like Adobe Photoshop, where we can copy and paste one image over an original one, it is simple to create fake images. Social media businesses’ top issue is doctored photographs. These altered photographs are the main source of false information and are frequently used to incite mobs. Since it is getting harder for the average person to tell the difference between a true image and one that has been altered thanks to the development of photo and video editing tools, forensic experts are needed. Social media is a virtual environment where many photographs are shared. There is no doubt that there are a lot of fake photographs with the numerous altering programmes available. By employing Error Level Analysis to forensically examine the image, it is possible to compare the actual and false photos’ levels of compression because they differ. While it is possible to edit the metadata, it is also possible to analyse the image’s metadata in order to determine whether it is real or phoney. It can be done by using a simple neural network with two convolutional layers, a max pooling layer, a dropout layer, two dense layers, and one output layer to apply deep learning to recognise images of manipulations using a dataset of a fake image and original images.