Manifold topological deep learning for biomedical data

The study introduces manifold topological deep learning (MTDL) that extends topological deep learning to data on differentiable manifolds through de Rham-Hodge theory. MTDL outperforms competing methods on various biomedical imaging benchmarks.