Source code for kornia.feature.nms

from typing import Tuple, Union

import torch
import torch.nn as nn


[docs]class NonMaximaSuppression2d(nn.Module): r"""Applies non maxima suppression to filter. """ def __init__(self, kernel_size: Tuple[int, int]): super(NonMaximaSuppression2d, self).__init__() self.kernel_size: Tuple[int, int] = kernel_size self.padding: Tuple[int, int] = self._compute_zero_padding2d(kernel_size) self.max_pool2d = nn.MaxPool2d(kernel_size, stride=1, padding=self.padding) @staticmethod def _compute_zero_padding2d( kernel_size: Tuple[int, int]) -> Tuple[int, int]: assert isinstance(kernel_size, tuple), type(kernel_size) assert len(kernel_size) == 2, kernel_size def pad(x): return (x - 1) // 2 # zero padding function ky, kx = kernel_size # we assume a cubic kernel return (pad(ky), pad(kx)) def forward(self, x: torch.Tensor) -> torch.Tensor: # type: ignore assert len(x.shape) == 4, x.shape # find local maximum values x_max: Union[torch.Tensor, Tuple[torch.Tensor, torch.Tensor]] = \ self.max_pool2d(x) # create mask for maximums in the original map x_mask: torch.Tensor = torch.where( x == x_max, torch.ones_like(x), torch.zeros_like(x)) return x * x_mask # return original masked by local max
# functiona api
[docs]def non_maxima_suppression2d( input: torch.Tensor, kernel_size: Tuple[int, int]) -> torch.Tensor: r"""Applies non maxima suppression to filter. See :class:`~kornia.feature.NonMaximaSuppression2d` for details. """ return NonMaximaSuppression2d(kernel_size)(input)