Abstract:
Nowadays many fake images are expanded through digital media and many
newspapers. Images are often directed with the intent and purpose of benefiting one
party. In fact, the images are often seen as the reveal of a fact or reality, therefore, false
news or any form of printing that using images that have been manipulated or tempering
[8] in such a way have the ability and potential to misinform the larger ones. To
detection image falsification of a huge number of image data is required, and an
architectural model that can process severally pixel in the image. In adding with project,
effectiveness and adjustability in the training data is also required to support its usage
in daily life. The concept of error level analysis big data and machine learning is the
right solution to this type of problem. Therefore, with the Convolutional Neural
Network (CNN) [9] architecture that utilizes Error Level Analysis (ELA) [6], image
forgery detection can reach above 80% and convergence with only this time.