kaira.metrics.image.StructuralSimilarityIndexMeasure

Inheritance diagram of StructuralSimilarityIndexMeasure

Inheritance diagram for StructuralSimilarityIndexMeasure

class kaira.metrics.image.StructuralSimilarityIndexMeasure(data_range: float = 1.0, kernel_size: int = 11, sigma: float = 1.5, reduction: str | None = None, *args: Any, **kwargs: Any)[source]

Bases: BaseMetric

Structural Similarity Index Measure (SSIM) Module.

SSIM measures the perceptual difference between two similar images. Values range from 0 to 1, where 1 means perfect similarity. The metric considers luminance, contrast, and structure to better match human visual perception [Wang et al., 2004] [Brunet et al., 2011].

Methods

__init__

Initialize the SSIM module.

compute_with_stats

Compute SSIM with mean and standard deviation.

forward

Calculate SSIM between predicted and target images.
__init__(data_range: float = 1.0, kernel_size: int = 11, sigma: float = 1.5, reduction: str | None = None, *args: Any, **kwargs: Any) → None[source]

Initialize the SSIM module.

Parameters:

  • data_range (float) – Range of the input data (typically 1.0 or 255)

  • kernel_size (int) – Size of the Gaussian kernel

  • sigma (float) – Standard deviation of the Gaussian kernel

  • reduction (Optional[str]) – Reduction method. The underlying torchmetrics implementation requires reduction=None, so this parameter controls post-processing reduction.

  • *args – Variable length argument list passed to the base class and torchmetrics.

  • **kwargs – Arbitrary keyword arguments passed to the base class and torchmetrics.

forward(x: Tensor, y: Tensor, *args: Any, **kwargs: Any) → Tensor[source]

Calculate SSIM between predicted and target images.

Parameters:

  • x (Tensor) – Predicted images

  • y (Tensor) – Target images

  • *args – Variable length argument list (currently unused).

  • **kwargs – Arbitrary keyword arguments (currently unused).

Returns:

SSIM values for each sample or reduced according to reduction parameter

Return type:

Tensor

compute_with_stats(x: Tensor, y: Tensor, *args: Any, **kwargs: Any) → Tuple[Tensor, Tensor][source]

Compute SSIM with mean and standard deviation.

Parameters:

  • x (Tensor) – Predicted images

  • y (Tensor) – Target images

  • *args – Variable length argument list (currently unused).

  • **kwargs – Arbitrary keyword arguments (currently unused).

Returns:

Mean and standard deviation of SSIM values

Return type:

Tuple[Tensor, Tensor]