Visual-Friendly Concept Protection via Selective Adversarial Perturbations

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Mi, Xiaoyue, Tang, Fan, Wu, You, Cao, Juan, Li, Peng, Liu, Yang
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912759952703488
author Mi, Xiaoyue
Tang, Fan
Wu, You
Cao, Juan
Li, Peng
Liu, Yang
author_facet Mi, Xiaoyue
Tang, Fan
Wu, You
Cao, Juan
Li, Peng
Liu, Yang
contents Personalized concept generation by tuning diffusion models with a few images raises potential legal and ethical concerns regarding privacy and intellectual property rights. Researchers attempt to prevent malicious personalization using adversarial perturbations. However, previous efforts have mainly focused on the effectiveness of protection while neglecting the visibility of perturbations. They utilize global adversarial perturbations, which introduce noticeable alterations to original images and significantly degrade visual quality. In this work, we propose the Visual-Friendly Concept Protection (VCPro) framework, which prioritizes the protection of key concepts chosen by the image owner through adversarial perturbations with lower perceptibility. To ensure these perturbations are as inconspicuous as possible, we introduce a relaxed optimization objective to identify the least perceptible yet effective adversarial perturbations, solved using the Lagrangian multiplier method. Qualitative and quantitative experiments validate that VCPro achieves a better trade-off between the visibility of perturbations and protection effectiveness, effectively prioritizing the protection of target concepts in images with less perceptible perturbations.
format Preprint
id arxiv_https___arxiv_org_abs_2408_08518
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Visual-Friendly Concept Protection via Selective Adversarial Perturbations
Mi, Xiaoyue
Tang, Fan
Wu, You
Cao, Juan
Li, Peng
Liu, Yang
Computer Vision and Pattern Recognition
Personalized concept generation by tuning diffusion models with a few images raises potential legal and ethical concerns regarding privacy and intellectual property rights. Researchers attempt to prevent malicious personalization using adversarial perturbations. However, previous efforts have mainly focused on the effectiveness of protection while neglecting the visibility of perturbations. They utilize global adversarial perturbations, which introduce noticeable alterations to original images and significantly degrade visual quality. In this work, we propose the Visual-Friendly Concept Protection (VCPro) framework, which prioritizes the protection of key concepts chosen by the image owner through adversarial perturbations with lower perceptibility. To ensure these perturbations are as inconspicuous as possible, we introduce a relaxed optimization objective to identify the least perceptible yet effective adversarial perturbations, solved using the Lagrangian multiplier method. Qualitative and quantitative experiments validate that VCPro achieves a better trade-off between the visibility of perturbations and protection effectiveness, effectively prioritizing the protection of target concepts in images with less perceptible perturbations.
title Visual-Friendly Concept Protection via Selective Adversarial Perturbations
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2408.08518