A Novel Approach to Detect Face Fraud Detection Using Artificial Intelligence

Abstract

The main aim of this research is to identify and prevent fraudulent activities which can be achieved through AI related to facial recognition systems. Nowadays the usage of facial recognition systems is very high, and in the same way the scams by fraudsters are also increased in this research by using AI bots instead of humans. The main motive of this research is to identify the misuse of facial recognition technology. The proposed method using CNN (Convolutional Neural Network) protects the individual privacy of people and their data. It detects whether the character in the image is an AI made or real human. This helps to ensure that only authorized people can use their information. Some of the sectors which use facial recognition systems are security, law enforcement, financial services, education, government services, retail. if unauthorized people access the above-mentioned sectors, the result will be imperiling. This method takes an image as an input and then python is used to process the image, importing a CV library to do this job. Next, we use deep learning models in python to identify whether the character in the image is AI generated or real human. The Computational Intelligence and Photography Lab at Yonsei University assembled a publicly available dataset for this work. Images of both real and fake human faces can be found in the Yonsei University Computational Intelligence and Photography Lab's database. The performance of the proposed system is measured using accuracy, precision and sensitivity. Experimental results shows that CNN based face recognition system outperforms.

Topics & Concepts