Reliable classification in digital and physical worlds under active adversaries and prior ambiguity
ContributorsTaran, Olga
DirectorsVoloshynovskyy, Svyatoslav
Defense date2021-06-09
Abstract
Keywords
- Printable graphical codes
- Copy detection patterns
- Adversarial attacks
- Hand-crafted attacks
- Machine learning attacks
- Supervised classification
- Semi-supervised classification
- One-class classification
Affiliation entities
Research groups
Citation (ISO format)
TARAN, Olga. Reliable classification in digital and physical worlds under active adversaries and prior ambiguity. Doctoral Thesis, 2021. doi: 10.13097/archive-ouverte/unige:152982
Main files (1)
Thesis
Identifiers
- PID : unige:152982
- DOI : 10.13097/archive-ouverte/unige:152982
- URN : urn:nbn:ch:unige-1529828
- Thesis number : Sc. 5566