Privacy-Preserving Image Features via Adversarial Affine Subspace Embeddings

TL;DR

This paper introduces a novel privacy-preserving image feature embedding method that embeds features within an affine subspace to prevent sensitive information recovery, enhancing privacy in cloud-based vision applications.

Contribution

The work proposes a new affine subspace embedding technique for image features that maintains matching capabilities while significantly improving privacy protection.

Findings

Effective in preventing private information recovery

Maintains feature matching accuracy

Applicable to visual localization, mapping, and face authentication

Abstract

Many computer vision systems require users to upload image features to the cloud for processing and storage. These features can be exploited to recover sensitive information about the scene or subjects, e.g., by reconstructing the appearance of the original image. To address this privacy concern, we propose a new privacy-preserving feature representation. The core idea of our work is to drop constraints from each feature descriptor by embedding it within an affine subspace containing the original feature as well as adversarial feature samples. Feature matching on the privacy-preserving representation is enabled based on the notion of subspace-to-subspace distance. We experimentally demonstrate the effectiveness of our method and its high practical relevance for the applications of visual localization and mapping as well as face authentication. Compared to the original features, our approach makes it significantly more difficult for an adversary to recover private information.