PPAuth: A Privacy-Preserving Framework for Authentication of Digital Image
- Publisher:
- Springer
- Publication Type:
- Conference Proceeding
- Citation:
- Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2023, 13914 LNCS, pp. 188-199
- Issue Date:
- 2023-01-01
Closed Access
| Filename | Description | Size | |||
|---|---|---|---|---|---|
| 978-3-031-34671-2_14.pdf | Published version | 1.13 MB |
Copyright Clearance Process
- Recently Added
- In Progress
- Closed Access
This item is closed access and not available.
Deep learning has been widely applied in many computer vision applications with remarkable success. Most of these techniques improve extraction of features from the fetched data for instance, image attributes are extracted for classifying an image. These attributes are quite useful for many tasks such as forensic investigation but it creates a threat for the privacy of the image if these attributes are leaked. Forensic authentication of digital images and videos is a crucial process in forensic investigation as they provide direct evidence. Authentication is a forensic process that involves verifying the authenticity of a digital image. In this paper, we propose a privacy-preserving framework to authenticate an image. It defends against the adversary from reconstructing the image from the extracted features and minimizes the private attribute leakage inference from the extracted features of the image. The main classifier performs the binary classification and tells whether an image is forged or authentic with an accuracy of 99%. Also, the framework maintains a very low accuracy for the adversarial classifier which aims at inferring some private attributes or even reconstruction of the image from these leaked attributes.
Please use this identifier to cite or link to this item: