Spherical Feature Pyramid Networks For Semantic Segmentation

TL;DR

This paper introduces spherical feature pyramid networks that leverage hierarchical graph-based CNNs to improve semantic segmentation on spherical data, outperforming existing spherical UNet models with fewer parameters.

Contribution

The paper proposes spherical FPNs inspired by planar FPNs, demonstrating superior performance and efficiency over spherical UNet architectures for spherical image segmentation.

Findings

Achieved state-of-the-art mIOU of 48.75 on Stanford 2D-3D-S dataset.

Spherical FPNs outperform spherical UNets in accuracy.

Models use fewer parameters than previous methods.

Abstract

Semantic segmentation for spherical data is a challenging problem in machine learning since conventional planar approaches require projecting the spherical image to the Euclidean plane. Representing the signal on a fundamentally different topology introduces edges and distortions which impact network performance. Recently, graph-based approaches have bypassed these challenges to attain significant improvements by representing the signal on a spherical mesh. Current approaches to spherical segmentation exclusively use variants of the UNet architecture, meaning more successful planar architectures remain unexplored. Inspired by the success of feature pyramid networks (FPNs) in planar image segmentation, we leverage the pyramidal hierarchy of graph-based spherical CNNs to design spherical FPNs. Our spherical FPN models show consistent improvements over spherical UNets, whilst using fewer parameters. On the Stanford 2D-3D-S dataset, our models achieve state-of-the-art performance with an mIOU of 48.75, an improvement of 3.75 IoU points over the previous best spherical CNN.