3D Shape

3D shape research focuses on accurately representing and manipulating three-dimensional objects from various data sources, aiming for robust and efficient methods for reconstruction, generation, and manipulation. Current research emphasizes the development and application of deep learning models, including diffusion models, transformers, and implicit neural representations (like signed distance functions and Gaussian splatting), often incorporating techniques like point cloud processing and multi-view geometry. These advancements have significant implications for diverse fields, such as robotics, computer-aided design, medical imaging, and cultural heritage preservation, by enabling more accurate 3D modeling and analysis from limited or noisy data.