VME & CDSI
Dataset

VME & CDSI

The Vehicles in the Middle East (VME) and Car Detection in Satellite Imagery (CDSI) datasets were developed to support robust and geographically diverse vehicle detection from high-resolution satellite imagery. VME focuses on the Middle East, covering 54 cities across 12 countries with more than 100,000 annotated vehicles, while CDSI consolidates VME with several existing public datasets to create a larger benchmark with nearly 900,000 car instances. Together, the datasets support the development and evaluation of more generalizable vehicle detection models for applications including humanitarian response, mobility analysis, and population displacement monitoring.

Team

Related publications

Humanitarian AI2025

AI-Driven Disaster Response and Displacement Monitoring

Noora Al-Emadi, Muhammad Imran, Yin Yang, Ingmar Weber, Fabjan Lashi, Gaia Rigodanza, Ivana Hajžmanová, Ferda Ofli — Communications of the ACM (2025)
DOI
Humanitarian AI2025

VME: A Satellite Imagery Dataset and Benchmark for Detecting Vehicles in the Middle East and Beyond

Noora Al-Emadi, Ingmar Weber, Yin Yang, Ferda Ofli — Scientific Data (2025)
DOI

Connections

ResearchPeoplePublications
VME & CDSIHumanitarian AIDr. Ferda OfliDr. Noora M. Al-EmadiAI-Driven Disaster Re…VME: A Satellite Imag…