metrics.py File Reference

Pose estimation evaluation metrics. More...

Go to the source code of this file.

Namespaces
	metrics

Functions
def	metrics._oks_matrix (tf.Tensor kpt0, tf.Tensor kpt1, tf.Tensor area, tf.Tensor stddev, Optional[float] eps=1e-7)
def	metrics._pose_area_calculation (tf.Tensor tensor)
def	metrics._matching_predictions (tf.Tensor tensor, tf.Tensor thres)
def	metrics.single_pose_oks (tf.Tensor y_true, tf.Tensor y_pred)
def	metrics.single_pose_heatmaps_oks (tf.Tensor y_true, tf.Tensor y_pred)
def	metrics.multi_pose_oks_mAP (tf.Tensor y_true, tf.Tensor y_pred)
def	metrics._precision_recall (tf.Tensor tp, tf.Tensor conf, tf.Tensor nb_gt, tf.Tensor maskpad, Optional[float] eps=1e-7)
def	metrics._auc (tf.Tensor precision, tf.Tensor recall)
def	metrics.compute_ap (tf.Tensor tp, tf.Tensor conf, tf.Tensor nb_gt, tf.Tensor maskpad, bool plot_metrics)

Detailed Description

Pose estimation evaluation metrics.

Implements OKS (Object Keypoint Similarity), AP (Average Precision), and AR (Average Recall) following the COCO evaluation protocol.

Note

Politecnico di Milano, A.Y. 2024-2025. Multidisciplinary Project — Neural Network Deployment on STM32N6 NPU.

Definition in file metrics.py.