|
MayaFlux 1.0.0
Digital-First Multimedia Processing Framework
|
| std::vector<BoundingBox> MayaFlux::Kinesis::Vision::VisionResult::component_boxes |
One BoundingBox per component from a ConnectedComponents step in this sequence, label 1..count, or empty.
Populated whenever ConnectedComponents ran, whether or not FindContours immediately follows: the box computation is a small readback of already GPU-reduced min/max buffers, not per-pixel work, so it costs nothing extra to keep regardless of what the sequence does next. label_id on each box matches Contour::label_id from a following FindContours step exactly, letting a caller join a contour's polygon to its own bounding box without re-deriving one from the polygon.
Definition at line 173 of file VisionContext.hpp.
Referenced by MayaFlux::Yantra::VisionGpuExecutor::op_connected_components().