Deep Learning model-based Multimedia forgery detection
Yash Shah, Parth Shah, Mansi Patel, Chinmay Khamkar, Pratik Kanani
Abstract
Images and videos can be spread very conveniently using social media platforms like WhatsApp and Facebook. The authenticity of this information cannot be verified easily but it spreads swiftly. Fake images or videos are a new threat for people as they spread false information and rumors. Advances in technology have given rise to several techniques that can easily generate fake images or videos. Deepfakes and spliced images are some of the results of such advances. They pose a great menace to the internet Tackling and detecting such an entity is a tricky task. Our paper portrays a technique to detect such entities. It will assist people in detecting bogus content and have confidence on the legitimacy of the content on the internet We present a description of CNN based approach and evaluate its results. The drawbacks of the traditional approach have been minimized using Inception Residual Network architecture based CNNs.