THE DETECTION OF VIDEO MANIPULATION OF FACES USING A NETWORK OF CONVOLUTIONAL NEURAL NETWORKS

THE DETECTION OF VIDEO MANIPULATION OF FACES USING A NETWORK OF CONVOLUTIONAL NEURAL NETWORKS

V.VENKATA RAMANJANEYULU1

ASSISTANT PROFESSOR

DEPT.OF CSE

TEEGALA KRISHNA REDDY ENGINEERING COLLEGE, MEERPET

P.MEENAKSHI2, P.MADHUSMITA3, PASAM BHARGAVA SAI4

DEPT. OF CSE

TEEGALA KRISHNA REDDY ENGINEERING COLLEGE, MEERPET

---------------------------------------------------------------------***---------------------------------------------------------------------

ABSTRACT

These days, it is not difficult to conceive scenarios that these lifelike face swapped deep fakes are used to influence political circumstances, organise terrorist acts, or blackmail people. In this paper, we offer a novel approach based on deep learning that is able to efficiently differentiate between actual movies and fraudulent ones created by artificial intelligence. Our approach employs a Res-Next Convolution neural network to extract the frame-level characteristics, and these features are then utilised to train a Long Short Term Memory (LSTM) based Recurrent Neural Network (RNN) to classify if the video has been subjected to any form of manipulation or not, i.e. whether the video is a deep fake or true video. Our technique is evaluated using a large quantity of balanced and mixed data sets, which were created by combining a number of other publicly accessible data sets, such as Face-Forensic++, the Deepfake detection challenge, and Celeb-DF.

Keywords: LSTM,RNN, data-set, Res-Next Convolution neural network, deep learning, AI