Open Access
This repository ensures open access to research outputs including datasets and code for secure and resilient AI, making them available for reuse, verification, and further development.
| Dataset | Partner | Dataset source | Purpose | Rationale | Link |
|---|---|---|---|---|---|
| Multi-Altitude Aerial Vehicles | UCY-KIOS CoE | Created in the context of previous research projects in which UCY-KIOS CoE was coordinating/participating | This dataset will be used to train models capable of robustly detecting objects of interest at various altitudes on edge devices. | One month as of the date of collectionRe-using this dataset ensures access to diverse and real-world aerial scenarios. | https://zenodo.org/records/7736336 |
| Small Object Aerial Person Detection | UCY-KIOS CoE | Created in the context of previous research projects in which UCY-KIOS CoE was coordinating/participating | This dataset will aid in training models for robust object detection at various altitudes, with a focus on detecting small, hard-to-identify objects on edge devices. | Utilizing this dataset provides essential data for fine-tuning edge AI models requiring precision in object detection. | https://zenodo.org/records/7740081 |
| AIDERv2 (Aerial Image Dataset for Emergency Response Applications) | UCY-KIOS CoE | Created in the context of previous research projects in which UCY-KIOS CoE was coordinating/participating | This dataset will be applied to assess the suitability of different methods for detecting objects of interest and understand whether additional contextual data is needed. Also, it will be used to evaluate how adversarial attacks affect model performance qualitatively. | This dataset contains information for testing the resilience of AI models against adversarial conditions and understanding performance dependencies. | https://zenodo.org/records/10891054 |