A Facial Feature-Based Binary Classification Framework for Deepfake Identification

DOI

10.2991/978-94-6239-789-7_4How to use a DOI?

Keywords

Deepfake image; CNN; Deepfake detection

Abstract

Deepfake technology represents a growing threat within digital media, not only because of the risks of misinforming the public but also of privacy violations and security breaches. This research works on the binary classification of facial features to increase deepfake detection. Taking advantage of the inherent differences between real and fake faces, we presented an effective method using deep learning to identify minute changes in texture, lighting, and facial movements. Our technique is centered around the extraction of facial regions most likely to be altered (eyes, mouth, skin, etc.) through automation of image segmentation. The next step is to use multiple CNNs and dense layers to obtain a high-level feature representation and to categorize it into real or fake. Besides, our method utilizes data augmentation along with ensemble techniques to make it more resistant to changes in quality and other environmental factors. The accomplishment of the advanced performance model with minimum computing power use has proven the effectiveness of the proposed framework on the Kaggle dataset. In conclusion, our study points out binary classification of facial features could be a key method for identifying deepfakes in a manner that is accurate, reliable, and scalable. Thus, we are hopeful that it will lead to the practical application of media forensics and digital security.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Gaurav Aggarwal
AU  - Gaurav Sharma
AU  - Ashutosh Bhatt
AU  - Pooja Joshi
AU  - Parul Saini
PY  - 2026
DA  - 2026/10/08
TI  - A Facial Feature-Based Binary Classification Framework for Deepfake Identification
BT  - Proceedings of theInternational Conference on Innovation and Practices in Artificial Intelligence (ICIPAI 2026)
PB  - Atlantis Press
SP  - 33
EP  - 44
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-789-7_4
DO  - 10.2991/978-94-6239-789-7_4
ID  - Aggarwal2026
ER  -