Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Secure Visual Data Processing via Federated Learning

Pedro Santos, Tânia Carvalho, Filipe Magalhães, Luís Antunes

TL;DR

This work tackles privacy in large-scale visual data processing by proposing a framework that fuses object detection, federated learning, and anonymization. It trains a shared detector across distributed data sources using a Flower-based FL setup with YOLOv8 and applies Gaussian blur to detected sensitive regions, evaluated on Open Images-derived data. The study shows a modest accuracy trade-off in FL but substantial privacy gains, supported by systematic analyses of epochs, rounds, and aggregation methods. This framework enables privacy-preserving visual data management suitable for privacy-sensitive domains such as surveillance, healthcare, and cross-institutional analytics.

Abstract

As the demand for privacy in visual data management grows, safeguarding sensitive information has become a critical challenge. This paper addresses the need for privacy-preserving solutions in large-scale visual data processing by leveraging federated learning. Although there have been developments in this field, previous research has mainly focused on integrating object detection with either anonymization or federated learning. However, these pairs often fail to address complex privacy concerns. On the one hand, object detection with anonymization alone can be vulnerable to reverse techniques. On the other hand, federated learning may not provide sufficient privacy guarantees. Therefore, we propose a new approach that combines object detection, federated learning and anonymization. Combining these three components aims to offer a robust privacy protection strategy by addressing different vulnerabilities in visual data. Our solution is evaluated against traditional centralized models, showing that while there is a slight trade-off in accuracy, the privacy benefits are substantial, making it well-suited for privacy sensitive applications.

Secure Visual Data Processing via Federated Learning

TL;DR

This work tackles privacy in large-scale visual data processing by proposing a framework that fuses object detection, federated learning, and anonymization. It trains a shared detector across distributed data sources using a Flower-based FL setup with YOLOv8 and applies Gaussian blur to detected sensitive regions, evaluated on Open Images-derived data. The study shows a modest accuracy trade-off in FL but substantial privacy gains, supported by systematic analyses of epochs, rounds, and aggregation methods. This framework enables privacy-preserving visual data management suitable for privacy-sensitive domains such as surveillance, healthcare, and cross-institutional analytics.

Abstract

As the demand for privacy in visual data management grows, safeguarding sensitive information has become a critical challenge. This paper addresses the need for privacy-preserving solutions in large-scale visual data processing by leveraging federated learning. Although there have been developments in this field, previous research has mainly focused on integrating object detection with either anonymization or federated learning. However, these pairs often fail to address complex privacy concerns. On the one hand, object detection with anonymization alone can be vulnerable to reverse techniques. On the other hand, federated learning may not provide sufficient privacy guarantees. Therefore, we propose a new approach that combines object detection, federated learning and anonymization. Combining these three components aims to offer a robust privacy protection strategy by addressing different vulnerabilities in visual data. Our solution is evaluated against traditional centralized models, showing that while there is a slight trade-off in accuracy, the privacy benefits are substantial, making it well-suited for privacy sensitive applications.

Paper Structure

This paper contains 28 sections, 3 figures, 5 tables.

Table of Contents

  1. INTRODUCTION
  2. LITERATURE REVIEW
  3. Object Detection
  4. One-Stage.
  5. Two-Stage.
  6. Anonymization Techniques
  7. Federated Learning
  8. Horizontal Federated Learning.
  9. Vertical Federated Learning.
  10. Current Research Directions
  11. VISUAL DATA PROTECTION AND LABELLING
  12. Methodology
  13. Methods
  14. Data
  15. RESULTS
  16. ...and 13 more sections

Figures (3)

  • Figure 1: Methodology for sensitive data detection and anonymization using federated learning.
  • Figure 2: Train and validation loss from baseline YOLOv8.
  • Figure 3: A visual representation of the anonymization layer applied to an image. The green bounding boxes highlight the areas where anonymization was applied, specifically targeting license plates and faces.