This package primarily contains the Vision python object which is the interface between the vision_bridge_node and gizmo_brain, as well as logic and data structures for analyzing vision results.

The interface to the vision_bridge_node provides conveniencies for managing vision modules as well as parsing FrameResults messages. The object also contains a buffer of processed frame results and methods for operating upon it for moment related functions. This includes scoring and threshold checks for deciding when to start capture and whether to upload or discard a moment.


The interface with the vision_bridge_node is through ros communication methods. This includes:

It communicates with gizmo_brain through methods on the python object. The primary points of interest are:

Vision module management:

.req_mods("<requestor_tag>", [<requested_module_states>])

Featurized results buffer:



It is instantiated within gizmo_brain and communicates with the vision_bridge_node.