Metrics

Metrics submodule.

This module contains various metrics and a metrics tracking class.

`RunningPSNR`

Compute the running PSNR during validation step in training.

This class allows to compute the PSNR on the entire validation set one batch at the time.

Attributes:

Name	Type	Description
`N`	`int`	Number of elements seen so far during the epoch.
`mse_sum`	`float`	Running sum of the MSE over the N elements seen so far.
`max`	`float`	Running max value of the N target images seen so far.
`min`	`float`	Running min value of the N target images seen so far.

`get()`

Get the actual PSNR value given the running statistics.

Returns:

Type	Description
`Optional[Tensor]`	PSNR value.

`reset()`

Reset the running PSNR computation.

Usually called at the end of each epoch.

`update(rec, tar)`

Update the running PSNR statistics given a new batch.

Parameters:

Name	Type	Description	Default
`rec`	`Tensor`	Reconstructed batch.	required
`tar`	`Tensor`	Target batch.	required

`avg_psnr(target, prediction)`

Compute the average PSNR over a batch of images.

Parameters:

Name	Type	Description	Default
`target`	`ndarray`	Array of ground truth images, shape is (N, C, H, W).	required
`prediction`	`ndarray`	Array of predicted images, shape is (N, C, H, W).	required

Returns:

Type	Description
`float`	Average PSNR value over the batch.

`avg_range_inv_psnr(target, prediction)`

Compute the average range-invariant PSNR over a batch of images.

Parameters:

Name	Type	Description	Default
`target`	`ndarray`	Array of ground truth images, shape is (N, C, H, W).	required
`prediction`	`ndarray`	Array of predicted images, shape is (N, C, H, W).	required

Returns:

Type	Description
`float`	Average range-invariant PSNR value over the batch.

`avg_range_invariant_psnr(pred, target)`

Compute the average range-invariant PSNR.

Parameters:

Name	Type	Description	Default
`pred`	`ndarray`	Predicted images.	required
`target`	`ndarray`	Target images.	required

Returns:

Type	Description
`float`	Average range-invariant PSNR value.

`avg_ssim(target, prediction)`

Compute the average Structural Similarity (SSIM) over a batch of images.

Parameters:

Name	Type	Description	Default
`target`	`ndarray`	Array of ground truth images, shape is (N, C, H, W).	required
`prediction`	`ndarray`	Array of predicted images, shape is (N, C, H, W).	required

Returns:

Type	Description
`tuple[float, float]`	Mean and standard deviation of SSIM values over the batch.

`multiscale_ssim(gt_, pred_, range_invariant=True)`

Compute channel-wise multiscale SSIM for each channel.

It allows to use either standard multiscale SSIM or its range-invariant version.

NOTE: images fed to this function should have channels dimension as the last one.

TODO: do we want to allow this behavior? or we want the usual (N, C, H, W)?

Parameters:

Name	Type	Description	Default
`gt_`	`Union[ndarray, Tensor]`	Ground truth image with shape (N, H, W, C).	required
`pred_`	`Union[ndarray, Tensor]`	Predicted image with shape (N, H, W, C).	required
`range_invariant`	`bool`	Whether to use standard or range invariant multiscale SSIM.	`True`

Returns:

Type	Description
`list[float]`	List of SSIM values for each channel.

`psnr(gt, pred, data_range)`

Peak Signal to Noise Ratio.

This method calls skimage.metrics.peak_signal_noise_ratio. See: https://scikit-image.org/docs/dev/api/skimage.metrics.html.

NOTE: to avoid unwanted behaviors (e.g., data_range inferred from array dtype), the data_range parameter is mandatory.

Parameters:

Name	Type	Description	Default
`gt`	`ndarray`	Ground truth array.	required
`pred`	`ndarray`	Predicted array.	required
`data_range`	`float`	The images pixel range.	required

Returns:

Type	Description
`float`	PSNR value.

`scale_invariant_psnr(gt, pred)`

Scale invariant PSNR.

Parameters:

Name	Type	Description	Default
`gt`	`ndarray`	Ground truth image.	required
`pred`	`ndarray`	Predicted image.	required

Returns:

Type	Description
`Union[float, tensor]`	Scale invariant PSNR value.