The Illegal Waste Dumping Detection (IWDD) contest is an international competition organized to encourage participants to develop advanced methods for real-time illegal waste dumping detection in videos captured by fixed CCTV cameras, runnable on smart cameras or embedded systems. Illegal dumping poses a major global issue, burdening cities with cleanup costs, harming ecosystems, and threatening public health.
We provide the participants with the MIVIA-IWDD dataset: 200 positive samples (100 static + 100 dynamic) and 200 negative samples. Each positive video includes precise timestamp annotations for fine-grained temporal localization and early detection training.
The private test set includes both positive and negative samples. The winner will be selected by the highest F1-Score on the test set. To emphasize real-time applicability, we also evaluate detection speed and computational efficiency to identify methods suitable for resource-constrained surveillance environments while maintaining high accuracy.
Full details: https://mivia.unisa.it/iwddcontest2026/