Efficient Temporal Perception and Video Understanding

Understanding a video means following what changes over time, often across many minutes of footage. We develop efficient methods for selecting informative frames, locating events and reasoning about their temporal relationships, alongside approaches that turn visual information into useful audio descriptions.

Experts

Name	Working area(s)	Contact
Marcel Klemt M.Sc.	Efficient Video Representation Learning	marcel.klemt@tu-... S4/23 211
Mohamed Abdelsalam	Efficient Multimodal Models, Efficient Video Understanding	mohamed.abdelsalam@tu-... S4/23 217
Aditya Arora M.Sc.	Multimodal Generation, Diffusion Models	aditya.arora@tu-... S4/23 217
Akshita Gupta M.Sc.	Multimodal Reasoning and Generation	akshita.gupta@tu-... S4\|23 217
Prof. Dr. Anna Rohrbach	Multimodal Grounded Learning	anna.rohrbach@tu-... S4/23 205
Prof. Dr. Marcus Rohrbach	Multimodal Reliable AI	office@mai.tu-... S4/23 204