reusex::vision::common::object Namespace Reference

Classes
struct	Box
	Axis-aligned bounding box defined by (left, top, right, bottom) coordinates. More...
struct	PosePoint
	A single keypoint for pose estimation, with visibility flag. More...
struct	Pose
	A set of keypoints representing a human or object pose. More...
struct	Obb
	Oriented bounding box (OBB) parameterized by center, size, and rotation angle. More...
struct	SegmentMap
	A raw binary segmentation map stored as pinned host memory. More...
struct	Segmentation
	Instance segmentation result backed by an OpenCV mask. More...
struct	Depth
	Depth map result with per-pixel depth values. More...
struct	Track
	Multi-object tracking state for a single tracked instance. More...
struct	DetectionBox
	Universal detection result container that holds a bounding box plus optional enriched data (pose, OBB, segmentation, depth, tracking). More...

Typedefs
using	DetectionBoxArray = std::vector<DetectionBox>
	Convenience alias for a collection of DetectionBox results.

Enumerations
enum class	ObjectType {   unknown = -1 , position = 0 , pose = 1 , obb = 2 ,   segmentation = 3 , depth_anything = 4 , depth_pro = 5 , track = 6 ,   detection = 7 }
	Enumeration of supported detection object types. More...

Functions
DetectionBox	create_box (float left, float top, float right, float bottom, float score, int class_id, const std::string &class_name)
	Create a standard detection box.
DetectionBox	create_position_box (float left, float top, float right, float bottom, float score, int class_id, const std::string &class_name)
	Create a position/region-of-interest box (type POSITION).
DetectionBox	create_track_box (float left, float top, float right, float bottom, float score, int track_id, const std::string &class_name)
	Create a simple tracking box without additional payload.
DetectionBox	create_track_box (float left, float top, float right, float bottom, float score, int track_id, const std::string &class_name, const object::Pose &pose)
	Create a tracking box that also carries pose keypoints.
DetectionBox	create_track_box (float left, float top, float right, float bottom, float score, int track_id, const std::string &class_name, const object::Obb &obb)
	Create a tracking box that also carries an oriented bounding box.
DetectionBox	create_track_box (float left, float top, float right, float bottom, float score, int track_id, const std::string &class_name, const object::Segmentation &seg)
	Create a tracking box that also carries a segmentation mask.
DetectionBox	create_obb_box (float cx, float cy, float w, float h, float angle, float score, int class_id, const std::string &class_name)
	Create an oriented bounding box detection.
DetectionBox	create_pose_box (float left, float top, float right, float bottom, const std::vector< PosePoint > &pose_points, float score, int class_id, const std::string &class_name)
	Create a pose-estimation detection box.
DetectionBox	create_segmentation_box (float left, float top, float right, float bottom, const cv::Mat &mask, float score, int class_id, const std::string &class_name)
	Create a segmentation detection box with an associated mask.
DetectionBox	create_depth_pro_box (const cv::Mat &depth, float fog_data)
	Create a depth detection from DepthPro output.
DetectionBox	create_depth_anything_box (const cv::Mat &depth)
	Create a depth detection from DepthAnything output.
cv::Mat	segment_map_to_mat (const std::shared_ptr< SegmentMap > &map)
	Convert a SegmentMap to an OpenCV Mat (zero-copy header wrapper).

Typedef Documentation

DetectionBoxArray

using reusex::vision::common::object::DetectionBoxArray = std::vector<DetectionBox>

Convenience alias for a collection of DetectionBox results.

Definition at line 216 of file object.hpp.

Enumeration Type Documentation

ObjectType

enum class reusex::vision::common::object::ObjectType

strong

Enumeration of supported detection object types.

Identifies the kind of data stored in a DetectionBox.

Enumerator
unknown	Unknown or unclassified object type.
position	Simple position/region of interest.
pose	Human pose estimation result.
obb	Oriented bounding box.
segmentation	Instance segmentation mask.
depth_anything	Depth map from DepthAnything model.
depth_pro	Depth map from DepthPro model.
track	Multi-object tracking result.
detection	Standard bounding box detection.

Definition at line 18 of file object.hpp.

Function Documentation

create_box()

DetectionBox reusex::vision::common::object::create_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		float	score,
		int	class_id,
		const std::string &	class_name )

Create a standard detection box.

Parameters

left	Left edge of the bounding box.
top	Top edge of the bounding box.
right	Right edge of the bounding box.
bottom	Bottom edge of the bounding box.
score	Confidence score in [0, 1].
class_id	Numeric class index.
class_name	Human-readable class label.

Returns

A DetectionBox with type DETECTION.

References create_box().

Referenced by create_box().

create_depth_anything_box()

DetectionBox reusex::vision::common::object::create_depth_anything_box

(

const cv::Mat &

depth

)

Create a depth detection from DepthAnything output.

Parameters

depth

CV_32F depth map (will be cloned).

Returns

A DetectionBox with type DEPTH_ANYTHING.

References create_depth_anything_box().

Referenced by create_depth_anything_box().

create_depth_pro_box()

DetectionBox reusex::vision::common::object::create_depth_pro_box	(	const cv::Mat &	depth,
		float	fog_data )

Create a depth detection from DepthPro output.

Parameters

depth	CV_32F depth map (will be cloned).
fog_data	Estimated fog/haze value.

Returns

A DetectionBox with type DEPTH_PRO.

References create_depth_pro_box().

Referenced by create_depth_pro_box().

create_obb_box()

DetectionBox reusex::vision::common::object::create_obb_box	(	float	cx,
		float	cy,
		float	w,
		float	h,
		float	angle,
		float	score,
		int	class_id,
		const std::string &	class_name )

Create an oriented bounding box detection.

The axis-aligned bounding box is computed automatically from the OBB parameters.

Parameters

cx	Center x.
cy	Center y.
w	Width.
h	Height.
angle	Rotation angle in degrees.
score	Confidence score.
class_id	Class index.
class_name	Class label.

Returns

A DetectionBox with type OBB.

References create_obb_box().

Referenced by create_obb_box().

create_pose_box()

DetectionBox reusex::vision::common::object::create_pose_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		const std::vector< PosePoint > &	pose_points,
		float	score,
		int	class_id,
		const std::string &	class_name )

Create a pose-estimation detection box.

Parameters

left	Left edge.
top	Top edge.
right	Right edge.
bottom	Bottom edge.
pose_points	Ordered list of pose keypoints.
score	Confidence score.
class_id	Class index.
class_name	Class label.

Returns

A DetectionBox with type POSE.

References create_pose_box().

Referenced by create_pose_box().

create_position_box()

DetectionBox reusex::vision::common::object::create_position_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		float	score,
		int	class_id,
		const std::string &	class_name )

Create a position/region-of-interest box (type POSITION).

Parameters

left	Left edge.
top	Top edge.
right	Right edge.
bottom	Bottom edge.
score	Confidence score.
class_id	Class index.
class_name	Class label.

Returns

A DetectionBox with type POSITION.

References create_position_box().

Referenced by create_position_box().

create_segmentation_box()

DetectionBox reusex::vision::common::object::create_segmentation_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		const cv::Mat &	mask,
		float	score,
		int	class_id,
		const std::string &	class_name )

Create a segmentation detection box with an associated mask.

Parameters

left	Left edge.
top	Top edge.
right	Right edge.
bottom	Bottom edge.
mask	Segmentation mask (will be cloned).
score	Confidence score.
class_id	Class index.
class_name	Class label.

Returns

A DetectionBox with type SEGMENTATION.

References create_segmentation_box().

Referenced by create_segmentation_box().

create_track_box() [1/4]

DetectionBox reusex::vision::common::object::create_track_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		float	score,
		int	track_id,
		const std::string &	class_name )

Create a simple tracking box without additional payload.

Parameters

left	Left edge.
top	Top edge.
right	Right edge.
bottom	Bottom edge.
score	Confidence score.
track_id	Unique tracking identifier.
class_name	Class label.

Returns

A DetectionBox with type TRACK.

References create_track_box().

Referenced by create_track_box(), create_track_box(), create_track_box(), and create_track_box().

create_track_box() [2/4]

DetectionBox reusex::vision::common::object::create_track_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		float	score,
		int	track_id,
		const std::string &	class_name,
		const object::Obb &	obb )

Create a tracking box that also carries an oriented bounding box.

References create_track_box().

create_track_box() [3/4]

DetectionBox reusex::vision::common::object::create_track_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		float	score,
		int	track_id,
		const std::string &	class_name,
		const object::Pose &	pose )

Create a tracking box that also carries pose keypoints.

References create_track_box().

create_track_box() [4/4]

DetectionBox reusex::vision::common::object::create_track_box	(	float	left,
		float	top,
		float	right,
		float	bottom,
		float	score,
		int	track_id,
		const std::string &	class_name,
		const object::Segmentation &	seg )

Create a tracking box that also carries a segmentation mask.

References create_track_box().

segment_map_to_mat()

cv::Mat reusex::vision::common::object::segment_map_to_mat

(

const std::shared_ptr< SegmentMap > &

map

)

Convert a SegmentMap to an OpenCV Mat (zero-copy header wrapper).

Parameters

map	The shared SegmentMap to wrap.

Returns

A CV_8UC1 Mat pointing to the SegmentMap's data buffer.

References segment_map_to_mat().

Referenced by segment_map_to_mat().