NinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learning

TL;DR

This paper introduces NinjaDesc, a novel adversarial learning approach to create visual descriptors that protect image privacy by preventing reconstruction while preserving matching accuracy for applications like camera localization.

Contribution

It presents a new adversarial training framework that balances privacy preservation with performance in visual descriptor generation.

Findings

Significantly reduces image reconstruction quality from descriptors.

Maintains high matching accuracy and camera localization performance.

Demonstrates effectiveness across various experimental scenarios.

Abstract

In the light of recent analyses on privacy-concerning scene revelation from visual descriptors, we develop descriptors that conceal the input image content. In particular, we propose an adversarial learning framework for training visual descriptors that prevent image reconstruction, while maintaining the matching accuracy. We let a feature encoding network and image reconstruction network compete with each other, such that the feature encoder tries to impede the image reconstruction with its generated descriptors, while the reconstructor tries to recover the input image from the descriptors. The experimental results demonstrate that the visual descriptors obtained with our method significantly deteriorate the image reconstruction quality with minimal impact on correspondence matching and camera localization performance.