EfficientNet-Based Multi-Class Detection of Real, Deepfake, and Plastic Surgery Faces

Authors: Li Kun, Milena Radenkovic

Published: 2025-09-12 17:38:51+00:00

AI Summary

This paper proposes a deep learning-based system for multi-class detection of real, deepfake, and plastic surgery altered faces. It utilizes EfficientNet-B4, ResNet-50, and VGG-16 models for classification after pre-processing with MTCNN feature extraction and Azure Face API. The research aims to develop advanced forgery detection tools that are robust against various forms of facial manipulation.

Abstract

Currently, deep learning has been utilised to tackle several difficulties in our everyday lives. It not only exhibits progress in computer vision but also constitutes the foundation for several revolutionary technologies. Nonetheless, similar to all phenomena, the use of deep learning in diverse domains has produced a multifaceted interaction of advantages and disadvantages for human society. Deepfake technology has advanced, significantly impacting social life. However, developments in this technology can affect privacy, the reputations of prominent personalities, and national security via software development. It can produce indistinguishable counterfeit photographs and films, potentially impairing the functionality of facial recognition systems, so presenting a significant risk. The improper application of deepfake technology produces several detrimental effects on society. Face-swapping programs mislead users by altering persons' appearances or expressions to fulfil particular aims or to appropriate personal information. Deepfake technology permeates daily life through such techniques. Certain individuals endeavour to sabotage election campaigns or subvert prominent political figures by creating deceptive pictures to influence public perception, causing significant harm to a nation's political and economic structure.

Key findings

The EfficientNet-B4 model achieved the highest performance with 97% accuracy after just 5 epochs. In comparison, the VGG-16 model also reached 97% accuracy but required 100 epochs, while the ResNet-50 model achieved less than 94% accuracy after 100 epochs. This indicates EfficientNet-B4 is the most efficient and effective model for this multi-class detection task among the evaluated architectures.

Approach

The approach involves pre-processing video frames to extract facial features using MTCNN and Azure Face API. These extracted face images are then fed into various deep convolutional neural networks, specifically EfficientNet-B4, ResNet-50, and VGG-16, for multi-class classification to distinguish between real, deepfake, and plastic surgery altered faces.

Datasets

HDA Plastic Surgery Face Database, DeeperForensics-1.0

Model(s)

EfficientNet-B4, ResNet-50, VGG-16, MTCNN (for feature extraction)

Author countries

← Previous