Task Automatic Video-based Motion Analysis (Completed)
Last Published:  07/31/19 10:05:33 AM (Central)
Short Title: AVIMA
Responsible HRP Element: Space Human Factors and Habitability
Collaborating Org(s):
Other:
Funding Status: Completed - Task completed and produced a deliverable
Procurement Mechanism(s):
Solicited
Aims:

During Phase I, Vecna demonstrated the feasibility of Automatic Video-based Motion Analysis (AVIMA), a video-based human motion analysis and visualization tool that automatically processes video footage of humans without the use of specialized markers. AVIMA features key innovations such as a video analytics module, a human motion analysis module which automatically estimates and tracks the 3D human body pose from video footage across a variety of operating scenarios, and a video analytics module which automatically interprets the subjects’ movements with respect to their operational environment, objects in the environment, and other crewmates. This video analytics module also locates, annotates, and counts how many times the crew performed a specific action, traversed a particular path, or accessed a piece of hardware.  


Phase II work leveraged the Phase I preliminary designs and prototype results. The team explored machine learning algorithms to detect human presence in a scene and completed a product that will detect when a person enters a specified region.

 

Mitigation/Partial mitigation of the gap/risk: This research assisted in defining requirements for an advanced tool to efficiently analyze 2-dimensional video.

Mappings