:j ddlZddlmZddlZddlmZmZddededefdZded ed ed edef d Zdd ededede eeeffdZ dd eded ee e e e ffdedef dZ y)N)Union)Tensortensorx block_sizereturnc|j\}}}}|dkDrtd|dtj|dz }tjt |dz |dz |}tjt t|jj|j}tj|dz } tjt |dz |dz |} tjt t| jj| j} |dddddd|f|dddddd|dzfz jdj} |dddddd|f|dddddd|dzfz jdj} | |dddd| ddf|dddd| dzddfz jdjz } | |dddd| ddf|dddd| dzddfz jdjz } |||z zdz }||dz z|z }|||z zdz }||dz z|z }| ||zz} | ||zz} | | kDr5tj|tjt||z nd}|| | z zS)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 ValueErrortorcharangerrangelistsettolistsymmetric_differencepowsummathlog2min)rr_channelsheightwidthhh_bh_bcvv_bv_bcd_bd_bcn_hbn_hbcn_vbn_vbcts x/home/mcse/projects/srt_converter/srt-converter-venv/lib/python3.12/site-packages/torchmetrics/functional/image/psnrb.py _compute_befr+s(    !|XYaXbblmnn UQYA ,,uZ!^UQY C DC <<S_AA#**,OP QD VaZ A ,,uZ!^VaZD EC <<S_AA#**,OP QD Q1c\?Qq!Qa/0 0 5 5c : > > @C aAtm qAq$(!23 3 8 8 = A A CDAaClOa1cAgq 01 1 6 6s ; ? ? AACQq!T1} !Qq!"3 4 4 9 9# > B B DDD UZ' (1 ,D uqy !T )E FZ' (1 ,D fqj !T )E4$;CEEMDADt * #fe*< ==QRA d sum_squared_errorbefnum_obs data_rangecN||z |z}dtj|dz|z zS)zComputes peak signal-to-noise ratio. Args: sum_squared_error: Sum of square of errors over all observations bef: block effect num_obs: Number of predictions or observations data_range: the range of the data. )r log10)r-r.r/r0s r*_psnrb_computer5Ds2*G3c9  JM,==> >>r,predstargetctjtj||z d}t|j |j }t ||}|||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 r3)devicer)r rrrnumelr9r+)r6r7rr-r/r.s r* _psnrb_updater<WsQ %))EFNA">?V\\^FMM:G u 4C c7 **r,c,t|trQtj||d|d}tj||d|d}t |d|dz }nt t |}t |||\}}}t||||S)aComputes `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: ground truth signal data_range: the range of the data. If a tuple is provided then the range is calculated as the difference and input is clamped between the values. block_size: integer indication the block size Return: Tensor with PSNRB score Example: >>> from torch import rand >>> from torchmetrics.functional.image import peak_signal_noise_ratio_with_blocked_effect >>> preds = rand(1, 1, 28, 28) >>> target = rand(1, 1, 28, 28) >>> peak_signal_noise_ratio_with_blocked_effect(preds, target, data_range=1.0) tensor(7.8402) rr )rmaxr:) isinstancetupler clamprfloatr<r5)r6r7r0rdata_range_valr-r.r/s r*+peak_signal_noise_ratio_with_blocked_effectrDfs@*e$ Ez!}*Q-HVAJqMJ 1 1 =>j 12&3E6j&Y#sG +S'> JJr,)) rtypingrr rrintr+r5r@r<rBrDr,r*rIs  ,F,,F,^?? ?? ?  ?& + + +S +vW]_eOeIf +& (K (K (KeU5%<001(K (K  (Kr,