Saved in:
Bibliographic Details
Main Authors: Kusakabe, Takeru, Hirose, Yudai, Mukaida, Mashiho, Ono, Satoshi
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.24792
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915701287026688
author Kusakabe, Takeru
Hirose, Yudai
Mukaida, Mashiho
Ono, Satoshi
author_facet Kusakabe, Takeru
Hirose, Yudai
Mukaida, Mashiho
Ono, Satoshi
contents Deep neural networks (DNNs) remain vulnerable to adversarial attacks that cause misclassification when specific perturbations are added to input images. This vulnerability also threatens the reliability of DNN-based monocular depth estimation (MDE) models, making robustness enhancement a critical need in practical applications. To validate the vulnerability of DNN-based MDE models, this study proposes a projection-based adversarial attack method that projects perturbation light onto a target object. The proposed method employs physics-in-the-loop (PITL) optimization -- evaluating candidate solutions in actual environments to account for device specifications and disturbances -- and utilizes a distributed covariance matrix adaptation evolution strategy. Experiments confirmed that the proposed method successfully created adversarial examples that lead to depth misestimations, resulting in parts of objects disappearing from the target scene.
format Preprint
id arxiv_https___arxiv_org_abs_2512_24792
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Projection-based Adversarial Attack using Physics-in-the-Loop Optimization for Monocular Depth Estimation
Kusakabe, Takeru
Hirose, Yudai
Mukaida, Mashiho
Ono, Satoshi
Computer Vision and Pattern Recognition
Machine Learning
Neural and Evolutionary Computing
Deep neural networks (DNNs) remain vulnerable to adversarial attacks that cause misclassification when specific perturbations are added to input images. This vulnerability also threatens the reliability of DNN-based monocular depth estimation (MDE) models, making robustness enhancement a critical need in practical applications. To validate the vulnerability of DNN-based MDE models, this study proposes a projection-based adversarial attack method that projects perturbation light onto a target object. The proposed method employs physics-in-the-loop (PITL) optimization -- evaluating candidate solutions in actual environments to account for device specifications and disturbances -- and utilizes a distributed covariance matrix adaptation evolution strategy. Experiments confirmed that the proposed method successfully created adversarial examples that lead to depth misestimations, resulting in parts of objects disappearing from the target scene.
title Projection-based Adversarial Attack using Physics-in-the-Loop Optimization for Monocular Depth Estimation
topic Computer Vision and Pattern Recognition
Machine Learning
Neural and Evolutionary Computing
url https://arxiv.org/abs/2512.24792