Hybrid Identity Deepfakes: A Comprehensive Review of Multi-Person Facial Blending and Its Impact on Forensic Detection

DOI

10.2991/978-94-6239-610-4_26How to use a DOI?

Keywords

Deepfake detection; Deep Learning Models- GANs & VAEs; Face morphing; Multi-identity synthesis; Digital media forensics

Abstract

Deep learning techniques have brought a revolution to digital media synthesis that can generate hyper realistic manipulated content, called deepfakes. Earlier image manipulation techniques included face morphing and manual compositing operated through basic geometric and pixel level processes, which limited both visual realism and practical scalability. The development of deep generative models through VAEs and GANs led to a major shift towards using data to learn complex visual and temporal patterns. The integration of attention mechanisms with transformer architectures and diffusion-based models has led to better synthesis quality for images and videos and multimodal content, which has significantly reduced the difference between artificial and real media. Scientists have worked to develop forensic methods that detect differences between authentic data and manipulated content through the development of handcrafted feature-based analysis, which has progressed into deep neural models that detect spatial and temporal inconsistencies. The rapid technology has created an ongoing competition that reduces the power of conventional detection indicators while showing how models perform when operating in real-time environments. Deepfake systems use hybrid synthesis methods to create new identities through face swapping by combining different facial features from various people. The multiple identity deepfakes create problems for forensic models because their generated faces lack any connection to actual people while showing signs of digital distortion. The research evaluates multiple-person image blending through a classification system which includes four main approaches: morphing-based, latent-space-based, temporal-based based and part-based facial compositing. The research demonstrates how hybrid identities break down three essential identity verification methods, which include artefact-based, identity-consistency checks and temporal cues. The paper also sketches future research directions that focus on developing identity coherence models that are resistant to attacks.

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  - M. Ashwini
AU  - Sai Shivani
PY  - 2026
DA  - 2026/05/05
TI  - Hybrid Identity Deepfakes: A Comprehensive Review of Multi-Person Facial Blending and Its Impact on Forensic Detection
BT  - Proceedings of the First International Conference on Advances in Forensics and Cyber Technologies (ICFACT 2025)
PB  - Atlantis Press
SP  - 275
EP  - 286
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6239-610-4_26
DO  - 10.2991/978-94-6239-610-4_26
ID  - Ashwini2026
ER  -