Large Scale Reconstruction

Large-scale 3D reconstruction aims to create detailed, three-dimensional models of extensive environments from various data sources, such as images, LiDAR, and radio signals. Current research focuses on improving the accuracy, efficiency, and scalability of reconstruction methods, employing techniques like Gaussian splatting, neural radiance fields (NeRFs), and graph neural networks, often incorporating deep learning for feature extraction and optimization. These advancements are impacting fields like robotics, augmented reality, and computational imaging by enabling the creation of high-fidelity digital twins of complex real-world scenes for applications ranging from autonomous navigation to medical imaging.

Papers