The code for the papers I’ve co-authored is or will be open-sourced, unless restricted by company policies.
Floor-sp: Inverse cad for floorplans by sequential room-wise shortest path
VSE-Infty: Learning the best pooling strategy for visual semantic embedding
Heat: Holistic edge attention transformer for structured reconstruction
Polydiffuse: Polygonal shape reconstruction via guided set diffusion models
MVDiffusion: Enabling Holistic Multi-view Image Generation with Correspondence-Aware Diffusion
MapTracker: Tracking with Strided Memory Fusion for Consistent Vector HD Mapping
PuzzleFusion++: Auto-agglomerative 3D Fracture Assembly by Denoise and Verify
MEGA-Bench: Scaling Multimodal Evaluation to over 500 Real-World Tasks
CLiFT: Compressive Light-Field Tokens for Compute-Efficient and Adaptive Neural Rendering