UNO is a underwater images dataset allowing deep-learning networks to localize non-natural objects within underwater images. The dataset consists in 279 videos, 5930 frames, and 10809 labels.

UNO is a more consistent and balanced version of the TrashCan image dataset to evaluate models for detecting non-natural objects in the underwater environment. We propose a method to balance the number of annotations and images for cross-evaluation. We then compare the performance of a SOTA object detection model when using TrashCAN and UNO datasets. Additionally, we assess covariate shift by testing the model on an image dataset for real-world application. Experimental results show significantly better and more consistent performance using the UNO dataset.