Privacy protection vs. utility in visual data

Abstract

Ubiquitous and networked sensors impose a huge challenge for privacy protection which has become an emerging problem of modern society. Protecting the privacy of visual data is particularly important due to the omnipresence of cameras, and various protection mechanisms for captured images and videos have been proposed. This paper introduces an objective evaluation framework in order to assess such protection methods. Visual privacy protection is typically realised by obfuscating sensitive image regions which often results in some loss of utility. Our evaluation framework assesses the achieved privacy protection and utility by comparing the performance of standard computer vision tasks, such as object recognition, detection and tracking on protected and unprotected visual data. The proposed framework extends the traditional frame-by-frame evaluation approach by introducing two new approaches based on aggregated and fused frames. We demonstrate our framework on eight differently protected video-sets and measure the trade-off between the improved privacy protection due to obfuscating captured image data and the degraded utility of the visual data. Results provided by our objective evaluation method are compared with an available state-of-the-art subjective study of these eight protection techniques.

Privacy–Enhancing Face Biometrics: A Comprehensive Survey

“…To assess the efficiency of privacy enhancement, most of the existing work relies on automatic recognition techniques trained for extracting various attributes, such as identity, age, gender, ethnicity, or others, from the biometric data, e.g., [10], [20], [22], [33], [44], [45], [46], [69], [105], [106]. Experiments are typically conducted on both, the original and privacy enhanced data, and performance differences between the two groups of data are reported.…”

“…Masking techniques are in general easy to implement and are, hence, often used as baselines in various privacyrelated performance evaluations, e.g., [104], [105], [106].…”

“…Here, the similarity with the original data is used as a proxy for biometric utility. This strategy is most often seen with image-level privacy enhancing techniques applied to surveillance footage, where computationally simple techniques, such as filtering, are used for privacy enhancement [106]. It is less suitable for more complex techniques that rely, for example, on image synthesis or other generative approaches.…”

“…Erdélyi et al [106] described an objective evaluation framework for assessing visual privacy enhancing methods with video surveillance data. The framework defined performance scores and evaluation protocols that measure the privacy protection achieved as well as the preserved levels of data utility.…”

Vae/Wgan-Based Image Representation Learning For Pose-Preserving Seamless Identity Replacement In Facial Images

“…This will require recognition of occupant's activities, gestures, body pose, facial expressions, etc. While it can be accomplished using video cameras, their use will likely raise privacy concerns which can hinder further development [7]. One approach to address privacy concerns is to obscure visuallyidentifying information via explicit image manipulation, such as pixelization, blurring, cartooning, etc.…”