Shortcuts

JaccardIndex#

ignite.metrics.JaccardIndex(cm, ignore_index=None)[source]#

Calculates the Jaccard Index using ConfusionMatrix metric. Implementation is based on IoU().

J(A,B)=ABAB\text{J}(A, B) = \frac{ \lvert A \cap B \rvert }{ \lvert A \cup B \rvert }
Parameters
Returns

MetricsLambda

Return type

ignite.metrics.metrics_lambda.MetricsLambda

Examples

cm = ConfusionMatrix(num_classes=3)
metric = JaccardIndex(cm, ignore_index=0)
metric.attach(default_evaluator, 'jac')
y_true = torch.Tensor([0, 1, 0, 1, 2]).long()
y_pred = torch.Tensor([
    [0.0, 1.0, 0.0],
    [0.0, 1.0, 0.0],
    [1.0, 0.0, 0.0],
    [0.0, 1.0, 0.0],
    [0.0, 1.0, 0.0],
])
state = default_evaluator.run([[y_pred, y_true]])
print(state.metrics['jac'])
tensor([0.5000, 0.0000], dtype=torch.float64)