ShapeR: Robust Conditional 3D Shape Generation from Casual Captures

ShapeR introduces a novel approach for robust conditional 3D object shape generation from casually captured image sequences. It leverages multi-modal inputs like SLAM points, posed images, and VLM-generated captions, and employs robust training techniques to generate high-fidelity 3D shapes. The method significantly outperforms existing approaches in challenging real-world settings, with applications in AR/VR and 3D content creation.