MediaPipePoseLandmarkerBase.cs File Reference

Classes
class	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe.MediaPipePoseLandmarkerBase< TNet >
	Shared MediaPipe PoseLandmarker algorithm implementation for any IDnnInferenceNet backend. Reproduces the pose landmarking graph logic of MediaPipe PoseLandmarker on top of the OpenCV Net DNN module. More...
struct	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe.MediaPipePoseLandmarkerBase< TNet >.NormalizedRect
	Pose ROI looped back from the previous frame. Equivalent to a MediaPipe NormalizedRect. More...
class	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe.MediaPipePoseLandmarkerBase< TNet >.PoseDetectionData
	One pose detection after letterbox removal, expressed in coordinates relative to the input image. Corresponds to the LOCATION_DATA payload of the original Detection. More...
struct	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe.MediaPipePoseLandmarkerBase< TNet >.PoseDetectorGraphResult
	Bundled outputs of PoseDetectorGraph. Corresponds to DETECTIONS, POSE_RECTS, and EXPANDED_POSE_RECTS in the original pose_detector_graph.cc. More...
struct	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe.MediaPipePoseLandmarkerBase< TNet >.PoseLandmarkerEstimationData
	Packed result for one pose from Pose Landmarker. Memory-compatible with each row produced by PackResultsToMats. Order: 33 elements corresponding to the original NormalizedLandmark values (x, y, z, visibility, presence) followed by 33 elements corresponding to the original Landmark values with the same five components. The user-facing MediaPipe PoseLandmarker result does not expose a pose-level raw presence score, so this packed representation does not include it either; the threshold is applied inside the graph using the constructor's minPosePresenceConfidence. More...
struct	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe.MediaPipePoseLandmarkerBase< TNet >.PoseResult
	Intermediate result for one pose before packing. Filled by the lower-level subgraph implementations. More...
class	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe.MediaPipePoseLandmarkerBase< TNet >.PoseSegmentationVisualizationBuffers
	Holds reusable Mat buffers used when visualizing vertically stacked pose segmentation, avoiding per-frame allocation.

Namespaces
namespace	OpenCVForUnity
namespace	OpenCVForUnity.Extensions
namespace	OpenCVForUnity.Extensions.Worker
namespace	OpenCVForUnity.Extensions.Worker.DnnModule
namespace	OpenCVForUnity.Extensions.Worker.DnnModule.MediaPipe

Enumerations
enum	KeyPoint : byte
	Landmark indices for the 33-body-point pose topology used by MediaPipe PoseLandmarker. These values correspond to the ordering of the original pose landmark list. More...
enum	MediaPipePoseRunningMode : byte
	Execution modes compatible with the MediaPipe PoseLandmarker task. This enum corresponds to the task running mode that switches between per-image processing and stateful video processing. More...

Enumeration Type Documentation

KeyPoint

enum KeyPoint : byte

Landmark indices for the 33-body-point pose topology used by MediaPipe PoseLandmarker. These values correspond to the ordering of the original pose landmark list.

MediaPipePoseRunningMode

enum MediaPipePoseRunningMode : byte

Execution modes compatible with the MediaPipe PoseLandmarker task. This enum corresponds to the task running mode that switches between per-image processing and stateful video processing.