Privacy-Preserving Autoencoder for Collaborative Object Detection

Privacy is a crucial concern in collaborative machine vision where a part of a Deep Neural Network (DNN) model runs on the edge, and the rest is executed on the cloud. In such applications, the machine vision model does not need the exact visual content to perform its task. Taking advantage of this potential, private information could be removed from the data insofar as it does not significantly impair the accuracy of the machine vision system. In this paper, we present an autoencoder-style network integrated within an object detection pipeline, which generates a latent representation of the input image that preserves task-relevant information while removing private information. Our approach employs an adversarial training strategy that not only removes private information from the bottleneck of the autoencoder but also promotes improved compression efficiency for feature channels coded by conventional codecs like VVC-Intra. We assess the proposed system using a realistic evaluation framework for privacy, directly measuring face and license plate recognition accuracy. Experimental results show that our proposed method is able to reduce the bitrate significantly at the same object detection accuracy compared to coding the input images directly, while keeping the face and license plate recognition accuracy on the images recovered from the bottleneck features low, implying strong privacy protection. Our code is available at https://github.com/bardia-az/ppa-code.

I. Introduction

With recent advancements in Artificial Intelligence (AI) applications such as autonomous driving, visual surveillance, and Internet of Things (IoT), the volume of data being transmitted between “intelligent” edge devices (e.g., cameras) and the cloud-based services is rapidly increasing. In these applications, the inference is performed collaboratively between an edge device, which often has limited computational capabilities, and a more powerful computing entity commonly referred to as the cloud. Given the rapidly growing number of such machine-to-machine (M2M) connections [1], there is an urgent need for an efficient data compression methodology tailored to automated machine-based analysis. JPEG AI [2], [3] and MPEG-VCM [4] are two prominent standardization groups actively working in this area to address this pressing need for data compression for machines [5]. As part of our research objectives in this work, we also aim to improve compression efficiency for such collaborative M2M analysis systems.