Learning with Weak Annotations for Robust Maritime Obstacle Detection

TL;DR

This paper introduces a scaffolding learning regime (SLR) that trains maritime obstacle detection networks using weak annotations, significantly reducing labeling effort while achieving equal or better performance than models trained on dense labels.

Contribution

The novel SLR method leverages weak annotations to efficiently train segmentation networks, outperforming dense-label training and enhancing domain generalization.

Findings

SLR reduces labeling effort by a factor of twenty.

SLR-trained networks match or outperform dense-label models.

SLR improves domain generalization and enables low-effort domain adaptation.

Abstract

Robust maritime obstacle detection is critical for safe navigation of autonomous boats and timely collision avoidance. The current state-of-the-art is based on deep segmentation networks trained on large datasets. However, per-pixel ground truth labeling of such datasets is labor-intensive and expensive. We propose a new scaffolding learning regime (SLR) that leverages weak annotations consisting of water edges, the horizon location, and obstacle bounding boxes to train segmentation-based obstacle detection networks, thereby reducing the required ground truth labeling effort by a factor of twenty. SLR trains an initial model from weak annotations and then alternates between re-estimating the segmentation pseudo-labels and improving the network parameters. Experiments show that maritime obstacle segmentation networks trained using SLR on weak annotations not only match but outperform the same networks trained with dense ground truth labels, which is a remarkable result. In addition to the increased accuracy, SLR also increases domain generalization and can be used for domain adaptation with a low manual annotation load. The SLR code and pre-trained models are available at https://github.com/lojzezust/SLR .