iou: Intersection-over-Union (IoU, Jaccard Index)

In stschn/deepANN: Neural Network Toolbox

Intersection-over-Union (IoU, Jaccard Index)

Description

Intersection-over-Union (IoU, Jaccard Index)

Usage

iou(actuals, preds, smooth = 1)

Arguments

actuals	A multidimensional array of actual values.
preds	A multidimensional array of prediction values.
smooth	A smoothing factor to avoid division by zero.

Details

This metric is used for evaluation of the results within image segmentation. actuals as well as preds are binary encoded image data masks in form of a n-dimensional array, mainly a two-dimensional array with the dimensions height and width for every channel. A value of 1 indicates the background (e.g. white color), a value equal 0 indicates the object (e.g. black color).

Value

Intersection-over-Union (IoU, Jaccard Index).

See Also

Other Metrics: accuracy(), cross_entropy(), dice(), entropy(), erf(), erfc(), erfcinv(), erfinv(), gini_impurity(), huber_loss(), log_cosh_loss(), mae(), mape(), mse(), msle(), quantile_loss(), rmse(), rmsle(), rmspe(), sse(), stderror(), vc(), wape(), wmape()

stschn/deepANN documentation built on June 25, 2024, 7:27 a.m.