Real-Time Application Of Deepfake For De-Identification Privacy Preservation And Data Protection

Inuwa, Muhsin

Abstract

In an era marked by mounting concerns over data privacy and protection, conventional
regulatory measures have proven inadequate against cyber-attacks, complicating data sharing
for research and development. Meanwhile, traditional face de-identification methods often
result in the complete erasure of facial information, hampering facial behaviour analysis. This
thesis addresses these challenges by proposing a real-time deepfake deidentification for privacy
preservation and data protection. Leveraging a first-order motion model and Mediapipe model
of deidentification, the study investigates methods to accurately identify multiple faces within
a single image, crucial for comprehensive deepfake models. Three distinct models were
developed and tested to achieve deidentification of created deepfakes. Experimentation from
various angles revealed differing levels of success, with considerations such as processing
power, model openness, and training data quality influencing outcomes. Despite challenges,
the study demonstrates the feasibility of real-time deepfake technology for privacy preservation
and data protection. The proposed pipeline offers potential solutions to ethical concerns
associated with data sharing, with implications extending to healthcare, autonomous vehicles,
and unmanned aerial vehicle technology.

Citation

Inuwa, M. (2024). Real-Time Application Of Deepfake For De-Identification Privacy Preservation And Data Protection. (Thesis). University of Salford

• Thesis Type: Thesis
• Deposit Date: Mar 26, 2024
• Publicly Available Date: May 26, 2024
• Award Date: Apr 25, 2024