o yi @sddlmZddlZddlmZddlmZddlmZddlm Z ddl m Z ded ed eeeffd d Z dded ede dedd ef ddZ dded ede dedd ef ddZdS))TupleN)Tensor)Literal)universal_image_quality_index)_check_same_shape)reducepredstargetreturncCs`|j|jkrtd|jd|jdt||t|jdkr,td|jd|jd||fS)zUpdates and returns variables required to compute Spectral Distortion Index. Checks for same shape and type of the input tensors. Args: preds: Low resolution multispectral image target: High resolution fused image z>Expected `ms` and `fused` to have the same data type. Got ms: z and fused: .z@Expected `preds` and `target` to have BxCxHxW shape. Got preds: z and target: )dtype TypeErrorrlenshape ValueError)rr rZ/home/ubuntu/.local/lib/python3.10/site-packages/torchmetrics/functional/image/d_lambda.py!_spectral_distortion_index_updates  relementwise_meanp reduction)rsumnonec Cs`|jd}t||f}t||f}t|D]h}t||D]`}t|dd||dddddf|dd||dddddf|||f<|||f<t|dd||dddddf|dd||dddddf|||f<|||f<qqtt|||} |dkrt| d|} ntd||dt| d|} t| |S)aComputes Spectral Distortion Index (SpectralDistortionIndex_) Args: preds: Low resolution multispectral image target: High resolution fused image p: a parameter to emphasize large spectral difference reduction: a method to reduce metric score over labels. - ``'elementwise_mean'``: takes the mean (default) - ``'sum'``: takes the sum - ``'none'``: no reduction will be applied Example: >>> _ = torch.manual_seed(42) >>> preds = torch.rand([16, 3, 16, 16]) >>> target = torch.rand([16, 3, 16, 16]) >>> preds, target = _spectral_distortion_index_update(preds, target) >>> _spectral_distortion_index_compute(preds, target) tensor(0.0234) rNg?) rtorchzerosrangerpowabsrr) rr rrlengthm1m2krdiffoutputrrr"_spectral_distortion_index_compute/s  ^`& r'cCs>t|tr |dkrtd|dt||\}}t||||S)aCalculates `Spectral Distortion Index`_ (SpectralDistortionIndex_) also known as D_lambda that is used to compare the spectral distortion between two images. Args: preds: Low resolution multispectral image target: High resolution fused image p: Large spectral differences reduction: a method to reduce metric score over labels. - ``'elementwise_mean'``: takes the mean (default) - ``'sum'``: takes the sum - ``'none'``: no reduction will be applied Return: Tensor with SpectralDistortionIndex score Raises: TypeError: If ``preds`` and ``target`` don't have the same data type. ValueError: If ``preds`` and ``target`` don't have ``BxCxHxW shape``. ValueError: If ``p`` is not a positive integer. Example: >>> from torchmetrics.functional import spectral_distortion_index >>> _ = torch.manual_seed(42) >>> preds = torch.rand([16, 3, 16, 16]) >>> target = torch.rand([16, 3, 16, 16]) >>> spectral_distortion_index(preds, target) tensor(0.0234) rz.Expected `p` to be a positive integer. Got p: r ) isinstanceintrrr')rr rrrrrspectral_distortion_index[s&r*)rr)typingrrrtyping_extensionsr!torchmetrics.functional.image.uqirtorchmetrics.utilities.checksr"torchmetrics.utilities.distributedrrr)r'r*rrrrsD       /