compute_precision_recall: Compute precision and recall statistics between predicted...

In weecology/NeonTreeEvaluation_package: Evaluate Individual Tree Segmentation Algorithms against Benchmark Data

Compute precision and recall statistics between predicted tree boxes and ground truth data

Description

Compute precision and recall statistics between predicted tree boxes and ground truth data

Usage

compute_precision_recall(
  ground_truth,
  predictions,
  threshold = 0.4,
  summarize = TRUE
)

Arguments

ground_truth	A SpatialPolygonDataFrame of ground truth polygons.
predictions	A SpatialPolygonDataFrame of prediction polygons.
threshold	The intersection-over-union threshold for a prediction overlap with a ground truth to be considered a true positive.
summarize	Logical. If true, return the precision and recall for this dataset, if false, return a data frame of matched crowns and the IoU overlap with ground truth.

Details

The numeric identity of the crown is stored in a column named crown_id. An error will be raised if this column does not exist.

Value

The recall and precision scores for the plot.

weecology/NeonTreeEvaluation_package documentation built on Aug. 27, 2024, 10:53 a.m.