hYFSrSSKJr SSKrSSKJr SSKJr SSKrSSK r SSK J r J r JrJrJrJrJrJr SSKJr SSKJs Js Jr SSKJr SS KJr /S Qr"S S 5r "S S\ 5r!"SS\ 5r""SS\5r#"SS5r$\\SSj55r%SSjr&\\SSj55r'\\SSj55r(\\SSj55r)S Sjr*g)!aEFunctional implementations for domain adaptation image transformations. This module provides low-level functions and classes for performing domain adaptation between images. It includes implementations for histogram matching, Fourier domain adaptation, and pixel distribution matching with various normalization techniques. ) annotationsN)deepcopy)Literal) add_weightedclipclipped from_floatget_num_channelspreserve_channel_dimto_floatuint8_io)Protocol)PCA)MONO_CHANNEL_DIMENSIONS)adapt_pixel_distributionapply_histogramfourier_domain_adaptationcF\rSrSrS SjrS SjrS SjrS SjrS SjrSr g) BaseScalercJSUlSUlSUlSUlSUlgN)data_mindata_maxmeanvarscale)selfs z/var/www/fran/franai/venv/lib/python3.13/site-packages/albumentations/augmentations/mixing/domain_adaptation_functional.py__init__BaseScaler.__init__s%+/ +/ '+ &*(, c[erNotImplementedErrorrxs rfitBaseScaler.fit&!!r"c[err$r&s r transformBaseScaler.transform)r*r"cFURU5 URU5$r)r(r,r&s r fit_transformBaseScaler.fit_transform,s  ~~a  r"c[err$r&s rinverse_transformBaseScaler.inverse_transform0r*r")rrrrrNreturnNoner' np.ndarrayr5r6r'r8r5r8) __name__ __module__ __qualname____firstlineno__r r(r,r/r2__static_attributes__r"rrrs-""!"r"rcN^\rSrSrSSU4SjjjrS SjrS SjrS SjrSrU=r $) MinMaxScaler4cX>[TU]5 USUlUSUlSUlg)Nr)superr minmax data_range)r feature_range __class__s rr MinMaxScaler.__init__5s- '*'*-1r"c[R"USS9Ul[R"USS9UlURUR- UlSUR UR S:H'gNraxisrD)nprFrrGrrHr&s rr(MinMaxScaler.fit;sQqq) qq) --$--7011,-r"cURbURb URc [S5e[R "XR5R [5n[R"X RUS9 [R"X RUR- US9 [R"X RUS9 U$)NpThis MinMaxScaler instance is not fitted yet. Call 'fit' with appropriate arguments before using this estimator.)out) rrrH ValueErrorrPsubtractastypefloatdividemultiplyrGrFaddrr'x_stds rr,MinMaxScaler.transformBs == DMM$9T__=TU   A}}-44U; %e4 EHHtxx/e< uhhE* r"cURbURb URc [S5eXR- UR UR- - R [5nX R-UR-$)NrS)rrrHrUrFrGrWrXr\s rr2MinMaxScaler.inverse_transformPsq == DMM$9T__=TU hh,488dhh#67??F&66r")rrrHrGrF))gg?)rIztuple[float, float]r5r6r7r9 r:r;r<r=r r(r,r2r> __classcell__rJs@rrArA4s!22 2 77r"rAcJ^\rSrSrSU4SjjrSSjrS SjrS SjrSrU=r $) StandardScalerZc">[TU]5 gr)rEr )rrJs rr StandardScaler.__init__[s r"c[R"USS9Ul[R"USS9Ul[R"UR5UlSURURS:H'grM)rPrrsqrtrr&s rr(StandardScaler.fit^sPGGAA& 66!!$WWTXX& &' 4::?#r"cURb URc [S5eXR- UR- $NzrThis StandardScaler instance is not fitted yet. Call 'fit' with appropriate arguments before using this estimator.rrrUr&s rr,StandardScaler.transformes? 99  2U II ++r"cURb URc [S5eXR-UR-$rmrnr&s rr2 StandardScaler.inverse_transformms? 99  2U JJ$))++r")rrrr4r7r9rarcs@rrereZs(,,,r"rec\rSrSr\R SSj5r\R SS Sjj5r\R SS Sjj5rSr g) TransformerInterfacevcgrr?r&s rr2&TransformerInterface.inverse_transformws>Ar"Ncgrr?rr'ys rr(TransformerInterface.fitzsMPr"cgrr?rxs rr,TransformerInterface.transform}sSVr"r?r9r)r'r8ryznp.ndarray | Noner5r8) r:r;r<r=abcabstractmethodr2r(r,r>r?r"rrsrsvs>AAPPVVr"rsct\rSrSrS S SjjrS SjrS SjrS SjrSSjr\ SSj5r SSjr S r g ) DomainAdaptercUuUlUl[U5UlXl[ U5UlURRURU55 gr) color_in color_outrsource_transformertarget_transformerr num_channelsr(flatten)r transformerref_imgcolor_conversionss rr DomainAdapter.__init__sO ):% t~"*;"7"-,W5 ##DLL$9:r"c`URcU$[R"XR5$r)rcv2cvtColorrimgs r to_colorspaceDomainAdapter.to_colorspaces#mm+sQc==1QQr"cURcU$[R"[U[R SS9UR5$)NTinplace)rrrrrPuint8rs rfrom_colorspaceDomainAdapter.from_colorspaces3 >> !J||Dbhh=t~~NNr"crURU5n[U5nURSUR5$)N)rr reshaperrs rrDomainAdapter.flattens2  %sm{{2t0011r"c[U[RSS9nURS:Xa UR UR X#55$UR UR X#UR55$)NTrrD)rrPrrrr)rpixelsheightwidths r reconstructDomainAdapter.reconstructs^fbhh5    !''v(EF F##FNN6$BSBS$TUUr"cj[R"[R"UR55$r)rPsigntrace components_)r's r _pca_signDomainAdapter._pca_signswwrxx .//r"c&URSSup#URU5nURRU5 [ UR S5(ar[ URS5(aWUR UR 5UR UR5:waUR =RS-slURRU5nUR RU5nURXbU5$)Nrr) shaperrr(hasattrrrrr,r2r)rimagerrrrepresentationresults r__call__DomainAdapter.__call__s BQ e$ ##F+ D++] ; ;//??t6674>>$JaJa;bb  # # / /2 5 /00::6B((::>J66r")rrrrrN))NN)rrsrr8rztuple[None, None])rr8r5r8)rr8rintrrr5r8r9)rr8r5r8) r:r;r<r=r rrrr staticmethodrrr>r?r"rrrsX 0< ;) ; ;- ;RO 2 V 007r"rc^URUR:wa [S5e[U5n[U5nXE:wa [S5eUS:Xa,[R"U5n[R"U5nUR UR :wa0[ R"XR SS[ RS9nURnU[R:Xa4[U[R5n[U[R5n[[[S.U"5n[XqS9nU"U5R![R5n UR![R5SU- -X--n U[R:XaU $[#U 5$) abAdapt the pixel distribution of an image to match a reference image. This function adapts the pixel distribution of an image to match a reference image using a specified transformation type and weight. Args: img (np.ndarray): The input image to be adapted. ref (np.ndarray): The reference image. transform_type (Literal["pca", "standard", "minmax"]): The type of transformation to use. weight (float): The weight of the transformation. Returns: np.ndarray: The adapted image. Raises: ValueError: If the input image and reference image have different dtypes or numbers of channels. z9Input image and reference image must have the same dtype.zFInput image and reference image must have the same number of channels.rDNr)dsize interpolation)pcastandardminmax)rr)dtyperUr rPsqueezerrresize INTER_AREAfloat32r rrrerArrWr ) rreftransform_typeweightimg_num_channelsref_num_channelsoriginal_dtyperadapter transformedrs rrrsB4 yyCIITUU',',+abb1jjojjo yyCIIjjIIbqMPYYN#bhh'bhh'>\RSabdK AG#,%%bjj1K ZZ #q6z 2[5I IF#rxx/6EXf5EEr"cURSSn[[R"[ U5U-55n[ R "U5upVUupx[S[Xd- 55[ [Xd-5U5p[S[XT- 55[ [XT-5U5pXU 2X24X U 2X24'U$)Nrr)rrrPfloorrF fgeometriccenterrG) amp_srcamp_trgbeta image_shapebordercenter_xcenter_yrrh1h2w1w2s rlow_freq_mutaters--#K #k*T12 3F#**;7HMF C)* +SX5F1G-P C)* +SX5F1G-O#rE25L1GrE25L Nr"c^UR[R5nUR[R5nUR[:Xa[R "USS9nUR[:Xa[R "USS9nUR Sn[R"U5n[U5GHgn[RRUSS2SS2U45n[RRUSS2SS2U45n [RRU5n [RRU 5n [R"U 5[R"U 5p[R"U 5n[U R5X5n[RR!U[R""SU -5-5n[RR%U5n[R&"U5USS2SS2U4'GMj U$)a$ Apply Fourier Domain Adaptation to the input image using a target image. This function performs domain adaptation in the frequency domain by modifying the amplitude spectrum of the source image based on the target image's amplitude spectrum. It preserves the phase information of the source image, which helps maintain its content while adapting its style to match the target image. Args: img (np.ndarray): The source image to be adapted. Can be grayscale or RGB. target_img (np.ndarray): The target image used as a reference for adaptation. Should have the same dimensions as the source image. beta (float): The adaptation strength, typically in the range [0, 1]. Higher values result in stronger adaptation towards the target image's style. Returns: np.ndarray: The adapted image with the same shape and type as the input image. Raises: ValueError: If the source and target images have different shapes. Note: - Both input images are converted to float32 for processing. - The function handles both grayscale (2D) and color (3D) images. - For grayscale images, an extra dimension is added to facilitate uniform processing. - The adaptation is performed channel-wise for color images. - The output is clipped to the valid range and preserves the original number of channels. The adaptation process involves the following steps for each channel: 1. Compute the 2D Fourier Transform of both source and target images. 2. Shift the zero frequency component to the center of the spectrum. 3. Extract amplitude and phase information from the source image's spectrum. 4. Mutate the source amplitude using the target amplitude and the beta parameter. 5. Combine the mutated amplitude with the original phase. 6. Perform the inverse Fourier Transform to obtain the adapted channel. The `low_freq_mutate` function (not shown here) is responsible for the actual amplitude mutation, focusing on low-frequency components which carry style information. Example: >>> import numpy as np >>> import albumentations as A >>> source_img = np.random.rand(100, 100, 3).astype(np.float32) >>> target_img = np.random.rand(100, 100, 3).astype(np.float32) >>> adapted_img = A.fourier_domain_adaptation(source_img, target_img, beta=0.5) >>> assert adapted_img.shape == source_img.shape References: FDA: Fourier Domain Adaptation for Semantic Segmentation: Yang and Soatto, 2020, CVPR https://openaccess.thecvf.com/content_CVPR_2020/papers/Yang_FDA_Fourier_Domain_Adaptation_for_Semantic_Segmentation_CVPR_2020_paper.pdf rrNNy?)rWrPrndimr expand_dimsr zeros_likerangefftfft2fftshiftabsanglercopy ifftshiftexpifft2real)r target_imgrsrc_imgtrg_imgr src_in_trg channel_idfft_srcfft_trgfft_src_shiftedfft_trg_shiftedrpha_srcr mutated_ampfft_src_mutatedsrc_in_trg_channels rrrsljj$G +G||....r2||....r2==$Lw'JL) &&++gaJ&678&&++gaJ&678&&//'2&&//'266/2BHH_4M&&)&gllngD &&**;W 9M+MN VV\\/:(*ww/A'B 1a#$/*2 r"c0URSSURSS:wa0[R"XRSURS4S9n[R"U5n[R"U5n[ X5n[ X2USU- 5$)a6Apply histogram matching to an input image using a reference image and blend the result. This function performs histogram matching between the input image and a reference image, then blends the result with the original input image based on the specified blend ratio. Args: img (np.ndarray): The input image to be transformed. Can be either grayscale or RGB. Supported dtypes: uint8, float32 (values should be in [0, 1] range). reference_image (np.ndarray): The reference image used for histogram matching. Should have the same number of channels as the input image. Supported dtypes: uint8, float32 (values should be in [0, 1] range). blend_ratio (float): The ratio for blending the matched image with the original image. Should be in the range [0, 1], where 0 means no change and 1 means full histogram matching. Returns: np.ndarray: The transformed image after histogram matching and blending. The output will have the same shape and dtype as the input image. Supported image types: - Grayscale images: 2D arrays - RGB images: 3D arrays with 3 channels - Multispectral images: 3D arrays with more than 3 channels Note: - If the input and reference images have different sizes, the reference image will be resized to match the input image's dimensions. - The function uses a custom implementation of histogram matching based on OpenCV and NumPy. - The @clipped and @preserve_channel_dim decorators ensure the output is within the valid range and maintains the original number of dimensions. NrrDr)r)rrrrPrmatch_histogramsr)rreference_image blend_ratiomatcheds rrr\sF yy!}--bq11**_YYq\399UV< U$)aAdjust an image so that its cumulative histogram matches that of another. The adjustment is applied separately for each channel. Args: image (np.ndarray): Input image. Can be gray-scale or in color. reference (np.ndarray): Image to match histogram of. Must have the same number of channels as image. channel_axis (int | None): If None, the image is assumed to be a grayscale (single channel) image. Otherwise, this indicates which axis of the array corresponds to channels. Returns: np.ndarray: Transformed input image. Raises: ValueError: Thrown when the number of channels in the input image and the reference differ. zrs1#  xxx&FF2H "",#7:#7L,Z,8W8W2727j 3F 3F 3F93F  3F  3F 3Fl  Z Zz +D +D\ & &RNr"