Clustering as Attention: Unified Image Segmentation with Hierarchical Clustering. (arXiv:2205.09949v1 [cs.CV])

We propose a hierarchical clustering-based image segmentation scheme for deep
neural networks, called HCFormer. We interpret image segmentation, including
semantic, instance, and panoptic segmentation, as a pixel clustering problem,
and accomplish it by bottom-up, hierarchical clustering with deep neural
networks. Our hierarchical clustering removes the pixel decoder from
conventional segmentation models and simplifies the segmentation pipeline,
resulting in improved segmentation accuracies and interpretability. HCFormer
can address semantic, instance, and panoptic segmentation with the same
architecture because the pixel clustering is …

arxiv attention clustering cv hierarchical image segmentation