Privacy-Preserving Clothing Classification using Vision Transformer for Thermal Comfort Estimation

TL;DR

This paper introduces a privacy-preserving clothing classification method using Vision Transformer that maintains high accuracy on encrypted images for thermal comfort estimation.

Contribution

The study presents a novel ViT-based classification scheme that preserves privacy without accuracy loss, unlike conventional pixel-based methods.

Findings

Conventional pixel-based methods suffer severe accuracy degradation on encrypted images.

The proposed ViT-based scheme maintains high accuracy on encrypted images.

The method effectively estimates clothing insulation for thermal comfort without compromising privacy.

Abstract

A privacy-preserving clothing classification scheme is presented to enable secure occupant-centric control (OCC) systems. Although the utilization of camera images for HVAC control has been widely studied to optimize thermal comfort, privacy protection of occupant images has not been considered in prior works. While various privacy-preserving methods have been proposed for image classification, applying conventional schemes results in severe accuracy degradation. In this paper, we introduce a privacy-preserving classification method using Vision Transformer (ViT) applied to clothing insulation estimation. In an experiment using the DeepFashion dataset categorized by clothing insulation, while the conventional pixel-based method suffers a severe accuracy drop, our scheme maintains a high accuracy on encrypted images, showing no degradation from plain images across all categories.