AI RESEARCH

GeodesicNVS: Probability Density Geodesic Flow Matching for Novel View Synthesis

arXiv CS.CV

ArXi:2603.01010v2 Announce Type: replace Recent advances in generative modeling have substantially enhanced novel view synthesis, yet maintaining consistency across viewpoints remains challenging. Diffusion-based models rely on stochastic noise-to-data transitions, which obscure deterministic structures and yield inconsistent view predictions. We advocate a Data-to-Data Flow Matching framework that learns deterministic transformations between paired views, enhancing view-consistent synthesis through explicit data coupling.