Hybrid Neural Representations for Spherical Data

In this paper, we study hybrid neural representations for spherical data, adomain of increasing relevance in scientific research. In particular, our workfocuses on weather and climate data as well as comic microwave background (CMB)data. Although previous studies have delved into coordinate-based neuralrepresentations for spherical signals, they often fail to capture the intricatedetails of highly nonlinear signals. To address this limitation, we introduce anovel approach named Hybrid Neural Representations for Spherical data (HNeR-S).Our main idea is to use spherical feature-grids to obtain positional featureswhich are combined with a multilayer perception to predict the target signal.We consider feature-grids with equirectangular and hierarchical equal areaisolatitude pixelization structures that align with weather data and CMB data,respectively. We extensively verify the effectiveness of our HNeR-S forregression, super-resolution, temporal interpolation, and compression tasks.