o oi @sJddlZdejdejdejfddZdejdejd ejdejfd d ZdS) Nsourcetemplatereturnc Cs|j}|}|}tj|ddd\}}}tj|dd\}}tj|d|jd}||d}tj|d|jd} | | d} t|| |} | ||S)aAdjust the pixel values of an image to match its histogram towards a target image. `Histogram matching `_ is the transformation of an image so that its histogram matches a specified histogram. In this implementation, the histogram is computed over the flattened image array. Code referred to `here `_. Args: source: Image to transform. template: Template image. It can have different dimensions to source. Returns: The transformed output image as the same shape as the source image. Note: This function does not matches histograms element-wisely if input a batched tensor. T)return_inverse return_counts)rr)dimdtype)shaperaveltorchuniquecumsumrinterpreshape) rroldshape_bin_idxs_countst_valuest_counts s_quantiles t_quantilesinterp_t_valuesrU/home/ubuntu/.local/lib/python3.10/site-packages/kornia/contrib/histogram_matching.pyhistogram_matchings   rxxpfpcCsfttj||dddt|d}||d||||||||d||||dS)aOne-dimensional linear interpolation for monotonically increasing sample points. Returns the one-dimensional piecewise linear interpolant to a function with given discrete data points :math:`(xp, fp)`, evaluated at :math:`x`. This is confirmed to be a correct implementation. See https://github.com/pytorch/pytorch/issues/1552#issuecomment-979998307 Args: x: the :math:`x`-coordinates at which to evaluate the interpolated values. xp: the :math:`x`-coordinates of the data points, must be increasing. fp: the :math:`y`-coordinates of the data points, same length as `xp`. Returns: the interpolated values, same size as `x`. T)right)r clip searchsortedlen)rrrirrrr@s"Dr)r Tensorrrrrrrs&+