Muhammad Saif Ullah Khan

(AG Augmented Vision, Prof. Dr. Didier Stricker)
hosted by PhD Presentation Day

"Motion Models for Robust Human Kinematics Capturing on Edge Devices"

The primary research objective is to modify and simplify vision-based neural networks to enable real-time detection of human motion on consumer devices like smartphones. The two main challenges are to run these networks efficiently on hardware with limited resources while maintaining or even improving their performance. This research has numerous applications in various industries, including sports, health, and entertainment. Currently, the most advanced models for these tasks are too large to be practically implemented. By optimizing these models, we can deploy them in real-world settings. Accurate movement analysis typically requires body-worn sensors, which restricts its use to labs. By improving these models for everyday devices, we aim to eliminate the need for such equipment, making motion capture technology more accessible and useful.

Time:	Monday, 29.01.2024, 15:00
Place:	48-680

Termin als Datei downloaden und in den Kalender importieren.