Projects one detection frame from its declared source coordinate space into
target.
Rectangles, polygons, polylines, and keypoints are scaled. Masks are left
untouched because a DetectionMask already carries the mask-pixel
dimensions its counts are encoded against; scaling them here would apply
the ratio twice. Frames without coordinateSpace, with a degenerate space,
or already in the target space are returned unchanged so mixed producers can
share one source.
Projects one detection frame from its declared source coordinate space into
target.Rectangles, polygons, polylines, and keypoints are scaled. Masks are left untouched because a
DetectionMaskalready carries the mask-pixel dimensions itscountsare encoded against; scaling them here would apply the ratio twice. Frames withoutcoordinateSpace, with a degenerate space, or already in the target space are returned unchanged so mixed producers can share one source.