Soft body pose-invariant evasion attacks against deep learning human detection

Simple item page
dc.contributor.authorLi, Chen
dc.contributor.authorGuo, Weisi
dc.date.accessioned2023-09-18T10:44:45Z
dc.date.available2023-09-18T10:44:45Z
dc.date.issued2023-09-01
dc.description.abstractEvasion attacks on deep neural networks (DNN) use manipulated data to let targets evade detection and/or classification across a wide range of DNNs. Most existing evasion attacks focus on planar images (e.g., photo, satellite imaging) and ignore the distortion of evasion in practical attacks (e.g., object rotation, deformation). Here, we build evasion patterns for soft-body human stakeholders, where patterns are designed to take into account body rotation, fabric stretch, printability, and lighting variations. We show that these are effective and robust to different human poses. This poses a significant threat to safety of autonomous vehicles and adversarial training should consider this new area.en_UK
dc.identifier.citationLi C, Guo W. (2023) Soft body pose-invariant evasion attacks against deep learning human detection. In: 2023 IEEE 9th International Conference on Big Data Computing Service and Applications (BigDataService), 17-20 July 2023, Athens, Greece. pp. 155-156en_UK
dc.identifier.isbn979-8-3503-3534-7
dc.identifier.urihttps://doi.org/10.1109/BigDataService58306.2023.00032
dc.identifier.urihttps://dspace.lib.cranfield.ac.uk/handle/1826/20229
dc.language.isoenen_UK
dc.publisherIEEEen_UK
dc.rightsAttribution 4.0 International*
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/*
dc.subjectDeep Learningen_UK
dc.subjectHuman Detectionen_UK
dc.subjectEvasion Attacken_UK
dc.subjectSafetyen_UK
dc.subjectSecurityen_UK
dc.titleSoft body pose-invariant evasion attacks against deep learning human detectionen_UK
dc.typeConference paperen_UK

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Soft_body_pose-invariant_evasion_attacks-2023.pdf
Size:
415.87 KB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.63 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Staff publications (SATM)