StegaVAR: Privacy-Preserving Video Action Recognition via Steganographic Domain Analysis

TL;DR

StegaVAR introduces a novel privacy-preserving video action recognition framework that embeds videos into cover videos and performs analysis within the steganographic domain, maintaining data privacy and accuracy.

Contribution

It is the first to perform VAR directly in the steganographic domain, addressing concealment and spatiotemporal disruption issues of prior methods.

Findings

Achieves superior VAR accuracy on benchmark datasets.

Effectively preserves privacy during transmission and analysis.

Compatible with multiple steganographic models.

Abstract

Despite the rapid progress of deep learning in video action recognition (VAR) in recent years, privacy leakage in videos remains a critical concern. Current state-of-the-art privacy-preserving methods often rely on anonymization. These methods suffer from (1) low concealment, where producing visually distorted videos that attract attackers' attention during transmission, and (2) spatiotemporal disruption, where degrading essential spatiotemporal features for accurate VAR. To address these issues, we propose StegaVAR, a novel framework that embeds action videos into ordinary cover videos and directly performs VAR in the steganographic domain for the first time. Throughout both data transmission and action analysis, the spatiotemporal information of hidden secret video remains complete, while the natural appearance of cover videos ensures the concealment of transmission. Considering the difficulty of steganographic domain analysis, we propose Secret Spatio-Temporal Promotion (STeP) and Cross-Band Difference Attention (CroDA) for analysis within the steganographic domain. STeP uses the secret video to guide spatiotemporal feature extraction in the steganographic domain during training. CroDA suppresses cover interference by capturing cross-band semantic differences. Experiments demonstrate that StegaVAR achieves superior VAR and privacy-preserving performance on widely used datasets. Moreover, our framework is effective for multiple steganographic models.