| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
|
|
| from typing import Sequence, Tuple, Union |
|
|
| import numpy as np |
|
|
| __all__ = ["same_padding", "calculate_out_shape", "gaussian_1d"] |
|
|
|
|
| def same_padding( |
| kernel_size: Union[Sequence[int], int], dilation: Union[Sequence[int], int] = 1 |
| ) -> Union[Tuple[int, ...], int]: |
| """ |
| Return the padding value needed to ensure a convolution using the given kernel size produces an output of the same |
| shape as the input for a stride of 1, otherwise ensure a shape of the input divided by the stride rounded down. |
| |
| Raises: |
| NotImplementedError: When ``np.any((kernel_size - 1) * dilation % 2 == 1)``. |
| |
| """ |
|
|
| kernel_size_np = np.atleast_1d(kernel_size) |
| dilation_np = np.atleast_1d(dilation) |
|
|
| if np.any((kernel_size_np - 1) * dilation % 2 == 1): |
| raise NotImplementedError( |
| f"Same padding not available for kernel_size={kernel_size_np} and dilation={dilation_np}." |
| ) |
|
|
| padding_np = (kernel_size_np - 1) / 2 * dilation_np |
| padding = tuple(int(p) for p in padding_np) |
|
|
| return padding if len(padding) > 1 else padding[0] |
|
|
|
|
| def calculate_out_shape( |
| in_shape: Union[Sequence[int], int], |
| kernel_size: Union[Sequence[int], int], |
| stride: Union[Sequence[int], int], |
| padding: Union[Sequence[int], int], |
| ) -> Union[Tuple[int, ...], int]: |
| """ |
| Calculate the output tensor shape when applying a convolution to a tensor of shape `inShape` with kernel size |
| `kernel_size`, stride value `stride`, and input padding value `padding`. All arguments can be scalars or multiple |
| values, return value is a scalar if all inputs are scalars. |
| """ |
| in_shape_np = np.atleast_1d(in_shape) |
| kernel_size_np = np.atleast_1d(kernel_size) |
| stride_np = np.atleast_1d(stride) |
| padding_np = np.atleast_1d(padding) |
|
|
| out_shape_np = ((in_shape_np - kernel_size_np + padding_np + padding_np) // stride_np) + 1 |
| out_shape = tuple(int(s) for s in out_shape_np) |
|
|
| return out_shape if len(out_shape) > 1 else out_shape[0] |
|
|
|
|
| def gaussian_1d(sigma: float, truncated: float = 4.0) -> np.ndarray: |
| """ |
| one dimensional gaussian kernel. |
| |
| Args: |
| sigma: std of the kernel |
| truncated: tail length |
| |
| Raises: |
| ValueError: When ``sigma`` is nonpositive. |
| |
| Returns: |
| 1D numpy array |
| |
| """ |
| if sigma <= 0: |
| raise ValueError(f"sigma must be positive, got {sigma}.") |
|
|
| tail = int(sigma * truncated + 0.5) |
| sigma2 = sigma * sigma |
| x = np.arange(-tail, tail + 1) |
| out = np.exp(-0.5 / sigma2 * x ** 2) |
| out /= out.sum() |
| return out |
|
|
