JaccardIndex

ignite.metrics.JaccardIndex(cm, ignore_index=None)

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

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

Parameters

• cm (ConfusionMatrix) – instance of confusion matrix metric

• ignore_index (Optional[int]) – index to ignore, e.g. background index

Returns

MetricsLambda

Return type

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)