Skip to main content

Video-Based Machine Learning for Action Recognition

Atomic action recognition focuses on simple, brief motion while complex actions last varying lengths from several seconds to several minutes. By using a combination of atomic actions and object data about the video, it may be possible to deduce the complex actions that the object of interest is taking. Our project explores this new concept and gathers the intermediate data of recognizing both atomic actions and objects so that it can be applied to determine complex actions in a video.

Team Members: 

Nicholas Edelman

Yue Shen

Mikalai Shlapkou

Irene Smith

Veronica Sun

Wenxuan Wu

Semester