The work performed in the video analysis activity aim to develop algorithms and software for the computerized annotation of the bee monitoring videos and for the automatic analysis of their behavior. These tools will be extended incrementally over the course of the project using the feedback from the biologist experts.
T2.1 Video annotation system
Objective: Enable the computerized annotation of the videos.
T2.2 Individual bee identification and tracking
Objective: Track and identify the individual bees from video.
T2.3 Behavior recognition using semi-supervised active learning
Objective: Enable the end-users to train custom behavior detectors capable of handling concepts such as fanning and bringing pollen back to the colony.
T2.4 Integrated HPC system
Objective: Prepare the algorithms and system for large-scale deployment on the High Performance Computing facility