a  d@sddlZddlmZddlZddlmZmZdeeedddZeeeeedd d Zdeeeeeeefd d d Z deeeed ddZ dS)N)Tuple)Tensortensor)x block_sizereturncCsb|j\}}}}|dkr&td|dt|d}tt|d|d|}ttt| |}t|d} tt|d|d|} ttt|  | } |dddddd|f|dddddd|df d } |dddddd|f|dddddd|df d } | |dddd| ddf|dddd| dddf d 7} | |dddd| ddf|dddd| dddf d 7} |||d}||d|}|||d}||d|}| ||} | ||} | | krRt |t t||nd}|| | S)zCompute block effect. Args: x: input image block_size: integer indication the block size Returns: Computed block effect Raises: ValueError: If the image is not a grayscale image z=`psnrb` metric expects grayscale images, but got images with z channels.Ng@r)shape ValueErrortorchZarangerrangelistsettolistsymmetric_differencepowsummathlog2min)rr_ZchannelsheightwidthhZh_bZh_bcvZv_bZv_bcZd_bZd_bcZn_hbZn_hbcZn_vbZn_vbctrl/var/www/html/stable-diffusion-webui/venv/lib/python3.9/site-packages/torchmetrics/functional/image/psnrb.py _compute_befs4  FFJJ  (r)sum_squared_errorbefn_obs data_rangercCs<|||}|dkr*dt|d|Sdtd|S)a-Computes peak signal-to-noise ratio. Args: sum_squared_error: Sum of square of errors over all observations bef: block effect n_obs: Number of predictions or observations data_range: the range of the data. If None, it is determined from the data (max - min).  g?)r log10)r r!r"r#rrr_psnrb_computeDs r')predstargetrrcCs>tt||d}t||jd}t||d}|||fS)zUpdates and returns variables required to compute peak signal-to-noise ratio. Args: preds: Predicted tensor target: Ground truth tensor block_size: Integer indication the block size r$)devicer)r rrrZnumelr*r)r(r)rr r"r!rrr _psnrb_updateYs  r,cCs2||}t|||d\}}}t||||S)a-Computes `Peak Signal to Noise Ratio With Blocked Effect` (PSNRB) metrics. .. math:: \text{PSNRB}(I, J) = 10 * \log_{10} \left(\frac{\max(I)^2}{\text{MSE}(I, J)-\text{B}(I, J)}\right) Where :math:`\text{MSE}` denotes the `mean-squared-error`_ function. Args: preds: estimated signal target: groun truth signal block_size: integer indication the block size Return: Tensor with PSNRB score Example: >>> import torch >>> from torchmetrics.functional.image import peak_signal_noise_ratio_with_blocked_effect >>> _ = torch.manual_seed(42) >>> preds = torch.rand(1, 1, 28, 28) >>> target = torch.rand(1, 1, 28, 28) >>> peak_signal_noise_ratio_with_blocked_effect(preds, target) tensor(7.8402) r+)maxrr,r')r(r)rr#r r!r"rrr+peak_signal_noise_ratio_with_blocked_effecthsr.)r)r)r) rtypingrr rrintrr'r,r.rrrrs$ 0