Intelligent robots rely on object detection models to perceive the environment. Following advances in deep learning security it has been revealed that object detection models are vulnerable to adversarial attacks. However, prior research primarily focuses on attacking static images or offline videos. Therefore, it is still unclear if such attacks could jeopardize real-world robotic applications in dynamic environments. This paper bridges this gap by presenting the first real-time online attack against object detection models. We devise three attacks that fabricate bounding boxes for nonexistent objects at desired locations. The attacks achieve a success rate of about 90% within about 20 iterations. The demo video is available at https://youtu.be/zJZ1aNlXsMU.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Adversarial Detection: Attacking Object Detection in Real Time


    Beteiligte:
    Wu, Han (Autor:in) / Yunas, Syed (Autor:in) / Rowlands, Sareh (Autor:in) / Ruan, Wenjie (Autor:in) / Wahlstrom, Johan (Autor:in)


    Erscheinungsdatum :

    04.06.2023


    Format / Umfang :

    3436652 byte





    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Adversarial Driving: Attacking End-to-End Autonomous Driving

    Wu, Han / Yunas, Syed / Rowlands, Sareh et al. | IEEE | 2023


    Attacking unreliability

    Knudsen, G.E. | Engineering Index Backfile | 1966


    ATTACKING UNRELIABILITY

    KNUDSEN, GEORGE | AIAA | 1966


    Attacking hybrid cost

    Brooke,L. | Kraftfahrwesen | 2006


    Attacking terminal problem

    Engineering Index Backfile | 1933