nicetoolbox.evaluation.metrics.point_cloud

Point cloud metrics for evaluating 3D point cloud data.

Classes

Jpe	Calculates the per joint position error for each frame.
PointCloudMetric	Handler for all point cloud metrics.

class nicetoolbox.evaluation.metrics.point_cloud.Jpe(**kwargs: Any)

Calculates the per joint position error for each frame.

Initializes the metric. Subclasses can use kwargs for specific setup.

compute() → Dict[Tuple[str, str, str], List[BatchResult] | AggregatedResult]

Compute the final metric from the stored state.

get_axis3(chunk: ChunkWorkItem) → List[str]

Get the joint names the metric is concerned with.

reset() → None

Reset the metric’s internal state.

update(preds: torch.Tensor, gts: torch.Tensor, meta_chunk: ChunkWorkItem, meta_frames: List[FrameInfo]) → None

Computes the per joint position error and stores it.

class nicetoolbox.evaluation.metrics.point_cloud.PointCloudMetric(cfg: EvaluationMetricType, device: str)

Handler for all point cloud metrics.

Initialize the metric handler with its config and device, creating its metrics.

Parameters:

• cfg (MetricTypeConfig) – Configuration for this metric type.

• device (str) – Device to run the metrics on (e.g., ‘cpu’ or ‘cuda’).

property name: str

The name of the metric handler (e.g., ‘point_cloud_metrics’).