Adversarial 3D Objects Against Monocular Depth Estimators

被引:0
|
作者
Feher, Tamas Mark [1 ]
Szemenyei, Marton [1 ]
机构
[1] Budapest Univ Technol & Econ, Dept Control Engn & Informat Technol, Budapest, Hungary
来源
PROCEEDINGS OF THE 2024 9TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING TECHNOLOGIES, ICMLT 2024 | 2024年
关键词
Monocular depth estimation; Adversarial attack; Transformer; Convolutional neural network; Implicit surfaces; ATTACKS;
D O I
10.1145/3674029.3674052
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The state-of-the-art solutions for the reconstruction of the depth from 2D images (monocular depth estimation) are deep neural networks. Although the neural networks in the field of computer vision generally provide good performance, they are quite sensitive to specially crafted inputs, called adversarial examples. It is possible to create these adversarial examples solely by manipulating the physical world with special lighting, painted textures, or 3D shapes. Unlike most of the existing literature in the field of monocular depth estimation, this paper focuses on adversarial 3D shapes instead of painted textures or lighting. According to our results, the adversarial 3D shapes can worsen the performance of monocular depth estimators, achieving an effect size comparable to the differences between subsequent state-of-the-art models. We published our implementation at: https://github.com/mntusr/threedattack1
引用
收藏
页码:138 / 142
页数:5
相关论文
共 50 条
  • [21] Categorical Depth Distribution Network for Monocular 3D Object Detection
    Reading, Cody
    Harakeh, Ali
    Chae, Julia
    Waslander, Steven L.
    2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 8551 - 8560
  • [22] Boosting Monocular Depth Estimation with Lightweight 3D Point Fusion
    Huynh, Lam
    Phong Nguyen
    Matas, Jiri
    Rahtu, Esa
    Heikkila, Janne
    2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, : 12747 - 12756
  • [23] 3D Packing for Self-Supervised Monocular Depth Estimation
    Guizilini, Vitor
    Ambrus, Rares
    Pillai, Sudeep
    Raventos, Allan
    Gaidon, Adrien
    2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2020, : 2482 - 2491
  • [24] PDR: Progressive Depth Regularization for Monocular 3D Object Detection
    Sheng, Hualian
    Cai, Sijia
    Zhao, Na
    Deng, Bing
    Zhao, Min-Jian
    Lee, Gim Hee
    IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, 2023, 33 (12) : 7591 - 7603
  • [25] Densely Constrained Depth Estimator for Monocular 3D Object Detection
    Li, Yingyan
    Chen, Yuntao
    He, Jiawei
    Zhang, Zhaoxiang
    COMPUTER VISION, ECCV 2022, PT IX, 2022, 13669 : 718 - 734
  • [26] Recognition of transparent objects Using 3D depth camera
    Yun, Youngjae
    Seo, Donghyeon
    Kim, Donghan
    2017 14TH INTERNATIONAL CONFERENCE ON UBIQUITOUS ROBOTS AND AMBIENT INTELLIGENCE (URAI), 2017, : 882 - 883
  • [27] Recognition of transparent objects using 3D depth camera
    Department of Electronics, Kyung Hee University, Yongin-si, Gyeonggi-do
    17104, Korea, Republic of
    Int. Conf. Ubiquitous Robot. Ambient Intell., URAI, 1600, (882-883):
  • [28] 3D measurements of micro-objects based on monocular wide-field optical microscopy with extended depth of field
    Qu, Yufu
    Zhang, Ping
    Hu, Yongbo
    MICROSCOPY RESEARCH AND TECHNIQUE, 2018, 81 (12) : 1434 - 1442
  • [29] Monocular Model-Based 3D Tracking of Rigid Objects: A Survey
    Lepetit, Vincent
    Fua, Pascal
    FOUNDATIONS AND TRENDS IN COMPUTER GRAPHICS AND VISION, 2005, 1 (01):
  • [30] eGAC3D: enhancing depth adaptive convolution and depth estimation for monocular 3D object pose detection
    Ngo, Duc Tuan
    Bui, Minh-Quan Viet
    Nguyen, Duc Dung
    Pham, Hoang-Anh
    PEERJ COMPUTER SCIENCE, 2022, 8