TrashCan: A Semantically-Segmented Dataset towards Visual Detection of Marine Debris
Jungseok Hong, Michael Fulton, Junaed Sattar
TL;DR
The paper tackles the challenge of autonomous marine debris detection by introducing TrashCan, a large underwater dataset with instance-segmentation annotations and two class configurations (TrashCan-Material and TrashCan-Instance) derived from JAMSTEC video frames. It establishes baselines using Faster R-CNN and Mask R-CNN with a ResNeXt-101/FPn backbone, converting annotations to COCO format and providing 7,212 labeled images for training and evaluation. The instance version generally yields better detection and segmentation performance, highlighting the value of fine-grained, instance-level labeling for underwater debris. By releasing TrashCan, the authors aim to accelerate research toward robust, onboard trash detection and removal for autonomous underwater systems.
Abstract
This paper presents TrashCan, a large dataset comprised of images of underwater trash collected from a variety of sources, annotated both using bounding boxes and segmentation labels, for development of robust detectors of marine debris. The dataset has two versions, TrashCan-Material and TrashCan-Instance, corresponding to different object class configurations. The eventual goal is to develop efficient and accurate trash detection methods suitable for onboard robot deployment. Along with information about the construction and sourcing of the TrashCan dataset, we present initial results of instance segmentation from Mask R-CNN and object detection from Faster R-CNN. These do not represent the best possible detection results but provides an initial baseline for future work in instance segmentation and object detection on the TrashCan dataset.