Development of a Privacy-Preserving UAV System With Deep Learning-Based Face Anonymization

Abstract

In this paper, we develop a privacy-preserving UAV system that does not infringe on the privacy of people in the videos taken by UAVs. Instead of blurring or masking the face parts of the videos, we want to exquisitely modify only the face parts so that the people in the modified videos still look like humans, but they become anonymous. Doing so, the semantic information of the videos can be preserved even with the anonymization. Specifically, based on the latest generative adversarial network architecture, we propose a deep learning-based face-anonymization scheme so that each modified face part looks like the face of a person who does not actually exist. The trained face-anonymizer is then mounted on the UAV system we have implemented. Through experiments, we confirm that the developed privacy-preserving UAV system anonymizes UAV's first-person videos so that the people in the video are not recognized as anyone in the dataset used. In addition, we show that even with such anonymized videos, the perception performance required for performing UAV's essential functions such as simultaneous localization and mapping is not degraded. INDEX TERMSPrivacy infringement, privacy-preserving vision, deep learning, security robot, UAV patrol system I. INTRODUCTION 1 As one of the most important systems in the 4-th industrial 2 era, unmanned aerial vehicles (UAVs) are expanding their 3 use in all directions, ranging from transportation, delivery, 4 surveillance, security, exploration, military, public safety, agriculture, and smart factories. In particular, UAV systems 6 capable of performing missions autonomously have bound-7 less potential in many applications. The recent rapid advance-8 ment of UAV systems is attributed to recent deep learning-9 based computer vision techniques. As UAV's cognitive abil-10 ity has soared, it has become possible to autonomously find 11 paths, avoid obstacles, and perform missions stably in various 12 situations. However, as UAVs become ubiquitous around us, 13 UAV's high-performance vision function may raise serious 14 concerns about privacy breaches by exposing us to unwanted 15

A Review on Security Issues and Solutions of the Internet of Drones

“…Lee et al [80] developed a privacy-preserving IoD system based on the state-of-the-art generative adversarial network architecture and deep learning techniques. Delicate modifications are made to only the face of individuals captured by drones such that they still look like human in the revised video, but are made anonymous.…”

Security Threats and Mitigation Techniques in UAV Communications: A Comprehensive Survey

“…The formulated problem was solved by dynamically adjusting the Lyapunov function and inducing a safe deep Q-learning network. A UAV-aided privacy-preserving system has been developed in [185] for anonymous masking people's faces in videos captured by UAVs without losing the semantic information.…”