Spherical Surface
Spherical surface research focuses on developing efficient and accurate methods for representing, analyzing, and processing data defined on spheres, a geometry prevalent in numerous scientific domains. Current research emphasizes novel algorithms for data fitting and interpolation on spherical surfaces, often leveraging techniques like spherical harmonics, convolutional neural networks tailored for spherical geometries (e.g., Spherical FNOs, HEAL-SWIN), and factorized attention mechanisms. These advancements are crucial for improving accuracy and efficiency in diverse applications, including weather forecasting, brain imaging analysis, and 3D object reconstruction, where spherical representations are essential.
Papers
HEAL-SWIN: A Vision Transformer On The Sphere
Oscar Carlsson, Jan E. Gerken, Hampus Linander, Heiner Spieß, Fredrik Ohlsson, Christoffer Petersson, Daniel Persson
Dynamic Object Tracking for Quadruped Manipulator with Spherical Image-Based Approach
Tianlin Zhang, Sikai Guo, Xiaogang Xiong, Wanlei Li, Zezheng Qi, Yunjiang Lou
Spherical Fourier Neural Operators: Learning Stable Dynamics on the Sphere
Boris Bonev, Thorsten Kurth, Christian Hundt, Jaideep Pathak, Maximilian Baust, Karthik Kashinath, Anima Anandkumar
Learning Representations on the Unit Sphere: Investigating Angular Gaussian and von Mises-Fisher Distributions for Online Continual Learning
Nicolas Michel, Giovanni Chierchia, Romain Negrel, Jean-François Bercher