An all-inclusive study on identification of fake multimedia content using Machine Learning approaches: A survey

Abstract: With the rapid progress of technologies like Artificial Intelligence (AI), deep learning, and Machine Learning(ML), the ability to manipulate and modify images has become more accessible. This has led to the emergence of a concerning phenomenon known as deepfakes, where criminals can generate deceptive videos, images, or audio content. In addressing this growing challenge, the present study introduces a comprehensive literature review on numerous approaches including Residual Networks (ResNet), Long Short-Term Memory (LSTM), and Convolutional Neural Networks (CNN) with reference to identification of fake content. The survey reviews popular research outcomes from the year 2017 to the current year. This work is useful in identifying research gaps in the area of fake content identification which can be useful to both experts and future enthusiastic. Despite being a preliminary study in this area, the authors propose this study as a valuable resource which can be used for years ahead.

Keywords: Deepfake, Deceptive content, ResNet, LSTM, and CNN.