Classification of real and fake human faces using deep learning for data security

Wan Suhaizal, Wan Muhammad Azim (2024) Classification of real and fake human faces using deep learning for data security. Project Report. Melaka, Malaysia, Universiti Teknikal Malaysia Melaka. (Submitted)

Text (Full Text) Classification of real and fake human faces using deep learning for data security.pdf - Submitted Version Download (1MB)

Abstract

The classification of real and fake human faces using deep learning has become increasingly critical for enhancing data security in various applications. This research focuses on developing robust deep learning models capable of accurately distinguishing between authentic and synthetic facial images. Leveraging techniques such as convolutional neural networks (CNNs) and advanced image processing algorithms, the study addresses the challenge of detecting sophisticated manipulations like deepfakes. A comprehensive dataset comprising diverse facial expressions, lighting conditions, and ethnicities is utilized for training and evaluating the models, ensuring their generalizability and reliability across different scenarios. The methodology includes image preprocessing, feature extraction, and model training using Google's Teachable Machine, facilitating intuitive model development and iteration. The effectiveness of the proposed approach is demonstrated through experimental evaluations, highlighting its potential to mitigate risks associated with fake media content and bolster data security measures. This research contributes to advancing the field of deep learning for facial recognition and underscores its role in safeguarding authenticity and trust in digital environments.

Actions (login required)

View Item