Naoya Muramatsu's Highlights on 'SC-DepthV3: Robust Self-supervised Monocular Depth Estimation for Dynamic Scenes' | Glasp