Supervized Segmentation with Graph-Structured Deep Metric Learning
This addresses segmentation challenges in domains like 3D point clouds, though it appears incremental as it builds on prior work.
The paper tackles the problem of segmenting graph-structured data by introducing a fully-supervised method using a graph-structured contrastive loss, which learns vertex embeddings to produce accurate segmentations, achieving state-of-the-art results on a 3D point cloud oversegmentation task.
We present a fully-supervized method for learning to segment data structured by an adjacency graph. We introduce the graph-structured contrastive loss, a loss function structured by a ground truth segmentation. It promotes learning vertex embeddings which are homogeneous within desired segments, and have high contrast at their interface. Thus, computing a piecewise-constant approximation of such embeddings produces a graph-partition close to the objective segmentation. This loss is fully backpropagable, which allows us to learn vertex embeddings with deep learning algorithms. We evaluate our methods on a 3D point cloud oversegmentation task, defining a new state-of-the-art by a large margin. These results are based on the published work of Landrieu and Boussaha 2019.