Fakultät für Informatik
Filtern
Erscheinungsjahr
Dokumenttyp
- Konferenzveröffentlichung (307)
- Wissenschaftlicher Artikel (230)
- Teil eines Buches (Kapitel) (44)
- Preprint (15)
- Bachelorarbeit (9)
- Bericht (9)
- Dissertation (7)
- Buch (Monographie) (5)
- Masterarbeit (2)
- Sonstiges (2)
Sprache
- Englisch (497)
- Deutsch (132)
- Mehrsprachig (2)
Schlagworte
- machine learning (20)
- process mining (20)
- business process management (12)
- security (12)
- eHealth (11)
- generative adversarial networks (11)
- digital innovation (10)
- Hochschuldidaktik (9)
- emotion recognition (9)
- design science research (8)
Institut
- Fakultät für Informatik (631)
- Fakultät für angewandte Geistes- und Naturwissenschaften (8)
- Fakultät für Elektrotechnik (7)
- Fakultät für Maschinenbau und Verfahrenstechnik (5)
- TTZ Donauwörth (4)
- Fakultät für Gestaltung (3)
- TTZ - Technologietransferzentrum Flexible Automation (Nördlingen) (3)
- School of Business (2)
- Fakultät für Architektur und Bauwesen (1)
Open Access
- ja (138)
- kostenlos im WWW (80)
- ja (Gold) (23)
- ja (Hybrid) (11)
- ja (Diamant) (4)
- ja (Grün) (2)
Begutachtungsstatus
- peer-reviewed (331)
- open peer reviewed (1)
- revised (1)
Lizenzen
- UrhG (420)
- CC-BY (133)
- CC BY-NC-ND (36)
- CC BY-NC (19)
- CC BY-SA (9)
- sonstiges (8)
- CC BY-NC-SA (5)
- CC BY-ND (1)
What If the External Crown Surface of Teeth Could Predict the Pulp Chamber? A DeepSDF-Based Approach
(2026)
HMI-Questionnaires: A Community-Driven, Open-Access Project for Human-Machine Interaction Scales
(2026)
Embedded Linux devices are susceptible to offline tampering and supply chain manipulation, rendering the integrity of persistent executables and shared libraries a practical security concern. This thesis integrates the Linux Integrity Subsystem into a Yocto-based embedded Linux image on a Raspberry Pi 4 and evaluates runtime integrity controls for persistent root filesystem artifacts.
A controlled configuration ladder is implemented, ranging from a baseline without integrity mechanisms to IMA measurement, IMA appraisal, and appraisal combined with EVM metadata protection. Two policy profiles are examined: a strict profile extending appraisal to file reads in addition to execution and executable memory mapping, and a minimal profile restricted to binaries and shared libraries.
Results indicate that policy scope dominates performance overhead. Under the strict profile, sequential and random read throughput decrease significantly, while boot time increases moderately. Furthermore, EVM prevents metadata forgery that can bypass hash-based appraisal.