Spatialcorrelationcoefficient

spatialcorrelationcoefficient

Metric SpatialCorrelationCoefficient from torchmetrics (torchmetrics.image.SpatialCorrelationCoefficient)

When to invoke this skill

The user has predictions + ground truth and asks to evaluate with SpatialCorrelationCoefficient, or mentions torchmetrics.image.SpatialCorrelationCoefficient directly, or wants the standard torchmetrics implementation.

Reference signature

from torchmetrics.image import SpatialCorrelationCoefficient

# SpatialCorrelationCoefficient(high_pass_filter: Optional[torch.Tensor] = None, window_size: int = 8, **kwargs: Any) -> None

Library docstring

Compute Spatial Correlation Coefficient (SCC_).

As input to ``forward`` and ``update`` the metric accepts the following input

- ``preds`` (:class:`~torch.Tensor`): Predictions from model of shape ``(N,C,H,W)`` or ``(N,H,W)``.
- ``target`` (:class:`~torch.Tensor`): Ground truth values of shape ``(N,C,H,W)`` or ``(N,H,W)``.

As output of `forward` and `compute` the metric returns the following output

- ``scc`` (:class:`~torch.Tensor`): Tensor with scc score

Args:
    hp_filter: High-pass filter tensor. default: tensor([[-1,-1,-1],[-1,8,-1],[-1,-1,-1]]).
    window_size: Local window size integer. default: 8.
    kwargs: Additional keyword arguments, see :ref:`Metric kwargs` for more info.

Example:
    >>> from torch import randn
    >>> from torchmetrics.image import SpatialCorrelationCoefficient as SCC
    >>> preds = randn([32, 3, 64, 64])
    >>> target = randn([32, 3, 64, 64])
    >>> scc = SCC()
    >>> scc(preds, target)
    tensor(0.0023)

Quick recipe

import torchmetrics.image as _m
score = _m.SpatialCorrelationCoefficient(y_true, y_pred)

Don'ts

• Don't reimplement when the library version handles edge cases (NaN, ties, empty inputs) better than a hand-rolled formula.
• Always check the library version's argument order — sklearn is (y_true, y_pred) while torchmetrics is (preds, target).