IWDD CONTEST 2027
SECOND EDITION
The WasteVision workshop welcomes scientific papers describing the methods and results developed within the Illegal Waste Dumping Detection Contest, an international competition devoted to advanced methods for real-time illegal waste dumping detection in videos acquired by fixed CCTV cameras and deployable on smart cameras or embedded systems.
Participants are granted access to the MIVIA-IWDD dataset, a perfectly balanced collection of 500 videos: 250 positive samples, evenly split between static and dynamic disposal, and 250 negative samples featuring challenging scenarios without any dumping event. Each positive video is annotated with the exact instant at which the illegal waste dumping occurs. Submitted methods are evaluated on a private test set and ranked according to the F1-Score, with GPU memory consumption adopted as a secondary criterion whenever two or more methods achieve comparable results.
The aim is to foster innovative solutions that remain reliable across real application scenarios of varying difficulty, promoting algorithms suitable for practical deployment in smart city surveillance systems.
DETECTION CHALLENGE CATEGORIES
STATIC DISPOSAL
Waste deliberately abandoned and left in unauthorized locations.
DYNAMIC DISPOSAL
Waste rapidly discarded while in motion, from vehicles or on foot.
