Point Label Aware Superpixels for Multi-species Segmentation of Underwater Imagery
This work solves the problem of efficient and accurate multi-species segmentation in underwater imagery for marine ecologists, representing an incremental improvement over existing label propagation and segmentation techniques.
The paper tackles the problem of automating coral reef monitoring by addressing the high cost of dense labeling for semantic segmentation models, proposing a point label aware superpixel method to propagate sparse labels and generate augmented ground truth. The result is improved performance, with increases of up to 14.32% in mean IoU on datasets and a 76% reduction in computation time compared to prior methods.
Monitoring coral reefs using underwater vehicles increases the range of marine surveys and availability of historical ecological data by collecting significant quantities of images. Analysis of this imagery can be automated using a model trained to perform semantic segmentation, however it is too costly and time-consuming to densely label images for training supervised models. In this letter, we leverage photo-quadrat imagery labeled by ecologists with sparse point labels. We propose a point label aware method for propagating labels within superpixel regions to obtain augmented ground truth for training a semantic segmentation model. Our point label aware superpixel method utilizes the sparse point labels, and clusters pixels using learned features to accurately generate single-species segments in cluttered, complex coral images. Our method outperforms prior methods on the UCSD Mosaics dataset by 3.62% for pixel accuracy and 8.35% for mean IoU for the label propagation task, while reducing computation time reported by previous approaches by 76%. We train a DeepLabv3+ architecture and outperform state-of-the-art for semantic segmentation by 2.91% for pixel accuracy and 9.65% for mean IoU on the UCSD Mosaics dataset and by 4.19% for pixel accuracy and 14.32% mean IoU for the Eilat dataset.