Marcus Rohrbach

University of California, Berkeley, Max Planck Institute for Informatics

Papers

Total Citations

161

H-Index

About

Marcus Rohrbach is a leading researcher in computer vision and multimodal machine learning, with a focus on video understanding, language grounding, and human-AI interaction. His most influential work, "The Long-Short Story of Movie Description" (2015, 124 citations), pioneered the integration of long-term temporal context with short-term visual cues for generating coherent natural language descriptions of complex video scenes. This contribution has profound implications for assistive technologies, such as aiding visually impaired individuals, and for advancing human-robot communication. Rohrbach also made early strides in 3D perception with his work on "3D Object Detection with Multiple Kinects" (2012, 37 citations), demonstrating robust multi-sensor fusion for real-world spatial understanding. Beyond these papers, his broader research spans visual question answering, video captioning, and learning from instructional videos, earning him recognition for bridging vision and language. With a citation impact that underscores his role in shaping modern video description tasks, Rohrbach’s work continues to inspire new directions in embodied AI and accessible technology.