Angel Chang

Angel Chang is a Hans Fischer Fellow at the Technical University of Munich (TUM), hosted by Prof. Matthias Nießner in the Visual Computing Focus Group. She holds affiliations with Simon Fraser University and Eloquent Labs. Her research bridges natural language processing with 3D scene understanding, focusing on grounding textual descriptions to 3D data through innovations like ScanNet and ShapeNet. Key contributions include dense captioning, object localization in RGB-D scans, and unified vision-language models.

Her work addresses challenges in 3D scene reconstruction, CAD model alignment, and interactive environments. Notable projects include SAPIEN for simulated part-based interactions and PartNet, a benchmark for hierarchical 3D object understanding. Chang's research aims to advance AI systems that interpret and generate complex 3D environments using multimodal data.

Publications span top venues like CVPR and ACM Transactions on Graphics, demonstrating impact in computer vision, robotics, and machine learning. She collaborates across institutions to develop datasets and frameworks that drive progress in digital scene modeling and human-AI collaboration.