Caller-supplied context for VisionIntent::TrackObjects. More...
#include <VisionInsight.hpp>
Collaboration diagram for MayaFlux::Kinesis::Vision::TrackObjectsContext:Public Attributes | |
| std::optional< BoundingBox > | subject_bounds |
| std::optional< uint32_t > | subject_track_id |
Detailed Description
Caller-supplied context for VisionIntent::TrackObjects.
subject_bounds narrows which tracked points count as the subject when more than one candidate is present in the raw TrackResult list, for the frame that first identifies a subject. subject_track_id continues an already-identified subject on later frames by its stable TrackResult::id instead of re-filtering by region every call, which is the case DetectFeaturesContext has no equivalent for: Keypoint carries no persistent identity at all, only TrackResult does. Not derived either way: the caller decides what the subject is, from whatever source (a previous frame's result, a UI selection, another sensor).
Definition at line 30 of file VisionInsight.hpp.
The documentation for this struct was generated from the following file:
