o oi~E@sddlmZmZmZddlZddlmZddlmZm Z m Z m Z m Z ddl mZddlmZmZddlmZdd lmZmZgd Z d+de de deeefdededede fddZ d+de deeefdededede f ddZ d,de de de dedededede fddZ d+de de deeefdededede fd d!Z   " #d-de de deeeefd$edeed%ed&ede fd'd(ZGd)d*d*eZ dS).)OptionalTupleUnionN)Resample)ModuleTensor as_tensorpadtensor)KORNIA_CHECK_SHAPE)infer_bbox_shape validate_bbox)resize)get_perspective_transform warp_affine) CenterCrop2D center_cropcrop_and_resize crop_by_boxescrop_by_indicescrop_by_transform_matbilinearzerosT input_tensorboxessizemode padding_mode align_cornersreturnc Cst|tstdt|t|tstdt|t|ttfs0t|dkr0td|t|jdkr@t d|jd|\}}| |}t ddg|d dg|d |d gd|d ggg|j |j d |jdd d } t||| |||S) aExtract crops from 2D images (4D tensor) and resize given a bounding box. Args: input_tensor: the 2D image tensor with shape (B, C, H, W). boxes : a tensor containing the coordinates of the bounding boxes to be extracted. The tensor must have the shape of Bx4x2, where each box is defined in the following (clockwise) order: top-left, top-right, bottom-right and bottom-left. The coordinates must be in the x, y order. The coordinates would compose a rectangle with a shape of (N1, N2). size: a tuple with the height and width that will be used to resize the extracted patches. mode: interpolation mode to calculate output values ``'bilinear'`` | ``'nearest'``. padding_mode: padding mode for outside grid values ``'zeros'`` | ``'border'`` | 'reflection'. align_corners: mode for grid_generation. Returns: Tensor: tensor containing the patches with shape BxCxN1xN2. Example: >>> input = torch.tensor([[[ ... [1., 2., 3., 4.], ... [5., 6., 7., 8.], ... [9., 10., 11., 12.], ... [13., 14., 15., 16.], ... ]]]) >>> boxes = torch.tensor([[ ... [1., 1.], ... [2., 1.], ... [2., 2.], ... [1., 2.], ... ]]) # 1x4x2 >>> crop_and_resize(input, boxes, (2, 2), mode='nearest', align_corners=True) tensor([[[[ 6., 7.], [10., 11.]]]]) 'Input tensor type is not a Tensor. Got z&Input boxes type is not a Tensor. Got 1Input size must be a tuple/list of length 2. Got 3Only tensor with shape (B, C, H, W) supported. Got .rrdevicedtype) isinstancer TypeErrortypetuplelistlen ValueErrorshapeAssertionErrortor r(r)expandr) rrrrrrdst_hdst_w points_src points_dstr:T/home/ubuntu/.local/lib/python3.10/site-packages/kornia/geometry/transform/crop2d.pyr(s$ -  ,rcCsPt|tstdt|t|ttfs"t|dkr"td|t|jdkr2t d|jd|\}}|jdd\}}|d} |d} |d} |d} | | } | | }| |d }||d }t | |g||g||g| |ggg|j |j d }t d d g|d d g|d |d gd |d ggg|j |j d  |jd d d }t||||||S) aCrop the 2D images (4D tensor) from the center. Args: input_tensor: the 2D image tensor with shape (B, C, H, W). size: a tuple with the expected height and width of the output patch. mode: interpolation mode to calculate output values ``'bilinear'`` | ``'nearest'``. padding_mode: padding mode for outside grid values ``'zeros'`` | ``'border'`` | ``'reflection'``. align_corners: mode for grid_generation. Returns: the output tensor with patches. Examples: >>> input = torch.tensor([[[ ... [1., 2., 3., 4.], ... [5., 6., 7., 8.], ... [9., 10., 11., 12.], ... [13., 14., 15., 16.], ... ]]]) >>> center_crop(input, (2, 4), mode='nearest', align_corners=True) tensor([[[[ 5., 6., 7., 8.], [ 9., 10., 11., 12.]]]]) r!r"r#r$r%r&Nrr'rr*)r+rr,r-r.r/r0r1r2r3r r(r)r5r)rrrrrr6r7src_hsrc_w dst_h_half dst_w_half src_h_half src_w_halfstart_xstart_yend_xend_yr8r9r:r:r;rss: "  ,rsrc_boxdst_boxvalidate_boxesc Cs|r t|t|t|jdkrtd|jdt||||}t|}|d|ddkrA|d|ddksPtd|dd|ddt|dd } t|dd } t ||| | f|||dS) a& Perform crop transform on 2D images (4D tensor) given two bounding boxes. Given an input tensor, this function selected the interested areas by the provided bounding boxes (src_box). Then the selected areas would be fitted into the targeted bounding boxes (dst_box) by a perspective transformation. So far, the ragged tensor is not supported by PyTorch right now. This function hereby requires the bounding boxes in a batch must be rectangles with same width and height. Args: input_tensor: the 2D image tensor with shape (B, C, H, W). src_box: a tensor with shape (B, 4, 2) containing the coordinates of the bounding boxes to be extracted. The tensor must have the shape of Bx4x2, where each box is defined in the clockwise order: top-left, top-right, bottom-right and bottom-left. The coordinates must be in x, y order. dst_box: a tensor with shape (B, 4, 2) containing the coordinates of the bounding boxes to be placed. The tensor must have the shape of Bx4x2, where each box is defined in the clockwise order: top-left, top-right, bottom-right and bottom-left. The coordinates must be in x, y order. mode: interpolation mode to calculate output values ``'bilinear'`` | ``'nearest'``. padding_mode: padding mode for outside grid values ``'zeros'`` | ``'border'`` | ``'reflection'``. align_corners: mode for grid_generation. validate_boxes: flag to perform validation on boxes. Returns: Tensor: the output tensor with patches. Examples: >>> input = torch.arange(16, dtype=torch.float32).reshape((1, 1, 4, 4)) >>> src_box = torch.tensor([[ ... [1., 1.], ... [2., 1.], ... [2., 2.], ... [1., 2.], ... ]]) # 1x4x2 >>> dst_box = torch.tensor([[ ... [0., 0.], ... [1., 0.], ... [1., 1.], ... [0., 1.], ... ]]) # 1x4x2 >>> crop_by_boxes(input, src_box, dst_box, align_corners=True) tensor([[[[ 5.0000, 6.0000], [ 9.0000, 10.0000]]]]) Note: If the src_box is smaller than dst_box, the following error will be thrown. RuntimeError: solve_cpu: For batch 0: U(2,2) is zero, singular U. r$r%r&rrzKCropping height, width and depth must be exact same in a batch. Got height z and width rrr) r r0r2r3rr4r allintitemr) rrGrHrrrrI dst_trans_srcbboxh_outw_outr:r:r;rs 90r transformout_sizecCsPt||jddd|j|jd}t||ddddddf||||d}|S)a~Perform crop transform on 2D images (4D tensor) given a perspective transformation matrix. Args: input_tensor: the 2D image tensor with shape (B, C, H, W). transform: a perspective transformation matrix with shape (B, 3, 3). out_size: size of the output image (height, width). mode: interpolation mode to calculate output values ``'bilinear'`` | ``'nearest'``. padding_mode (str): padding mode for outside grid values ``'zeros'`` | ``'border'`` | ``'reflection'``. align_corners: mode for grid_generation. Returns: the output tensor with patches. rr*r'Nr"rJ)rr5r2r(r)r)rrRrSrrrrNpatchesr:r:r;rs rFr interpolation antialiasshape_compensationc Csft|gdt|gd|j\}}} } t||jtjd} | ddddf} | ddddfd} | ddddf} | ddddfd}t| jdd t| jdd krrt| jdd krrt|jdd krrdkrnn2|d t| dt|dt| dt| df}|dur|jd d|krt ||||d |d S|durt | \}}|jdd |jdd f}tj ||g|R|j|j d}t |D]`\}} |||dddt| |t||t| |t| |f}|jd d|kr,|dkrt ||||d |d ||<qt|d|d|jdd|d|jd g||<q|||<q|S)a\Crop tensors with naive indices. Args: input_tensor: the 2D image tensor with shape (B, C, H, W). src_box: a tensor with shape (B, 4, 2) containing the coordinates of the bounding boxes to be extracted. The tensor must have the shape of Bx4x2, where each box is defined in the clockwise order: top-left, top-right, bottom-right and bottom-left. The coordinates must be in x, y order. size: output size. An auto resize or pad will be performed according to ``shape_compensation`` if the cropped slice sizes are not exactly align `size`. If None, will auto-infer from src_box. interpolation: algorithm used for upsampling: ``'nearest'`` | ``'linear'`` | ``'bilinear'`` | 'bicubic' | 'trilinear' | 'area'. align_corners: interpolation flag. antialias: if True, then image will be filtered with Gaussian before downscaling. No effect for upscaling. shape_compensation: if the cropped slice sizes are not exactly align `size`, the image can either be padded or resized. )BCHW)rX42r'NrrF)sorted.r<short)rUrsiderVrr*)r r2rr(torchlongr0uniquerLrr emptyr) enumerater )rrGrrUrrVrWrXrY_srcx1x2y1y2outhwi_outr:r:r;r<sR 6   D  4 rc sneZdZdZdejjdfdeee eeffde dee eefde dd f fd d Z d e de fd dZZS)raFCenter crop the input tensor. Args: size: Size (h, w) in pixels of the resized region or just one side. align_corners: interpolation flag. resample: Resampling mode. cropping_mode: Cropping mode, "resample" or "slice". Note: For JIT, the cropping mode must be "resample". Tslicerrresample cropping_moder Ncstt|tr|d|df|_nt|tr||f|_n tdt|d|j\}}tj ddg|ddg|d|dgd|dgggtj d}| d| | d|t ||||jdd |_dS) Nrrz8Invalid size type. Expected (int, tuple(int, int). Got: r&)r)r8r9r)rsrtrrr)super__init__r+r.rrL Exceptionr-rbr rcregister_bufferclonergetflags)selfrrrsrtr6r7r9 __class__r:r;rvs     :  zCenterCrop2D.__init__inputcCs|jd}|j\}}|jdd\}}|d|d}}|d|d} } t| |t| |} } | |d| |d} }| |jd<| |jd<| |jd<| |jd<| |jd <||jd <| |jd <||jd <|jd dkrt|j|dd||j|dd|}||dd}t ||ddddddf|j|jdj d|jdS|jd dkrt ||j|dd||jdSt d|jd d)Nrr<r"r)rrr)rrr)rrr)rrr)rr"r)rr"r)rr^r)rr^rrtrsr*rrrrrzNot supported type: r&)r2rrLr8r{rr5r4r9rnamelowerrNotImplementedError)r|r batch_sizer6r7r=r>r?r@rArBrCrDrErFrRr:r:r;forwards@           $zCenterCrop2D.forward)__name__ __module__ __qualname____doc__rBILINEARrrrLrboolstrrvrr __classcell__r:r:r}r;rs" r)rrT)rrTT)NrNFr)!typingrrrrbkornia.constantsr kornia.corerrrr r kornia.core.checkr kornia.geometry.bboxr r affwarprimgwarprr__all__rLrrrrrrrrr:r:r:r;s     N  R V  , K