gaussian

dnois.sensor.gaussian(signal: Tensor, sigma: float | tuple[float, float] | Tensor, noise_only: bool = False, dims: int | Sequence[int] = 0) → Tensor

Applying gaussian noise to signal:

\[\tilde{\mathbf{x}}=\mathbf{x}+\mathbf{n}, \mathbf{n}\sim\mathcal{N}(0,\sigma^2\mathbf{I}).\]
Parameters:

  • signal (Tensor) – Input signal.

  • sigma (float | tuple[float, float] | Tensor) – Standard deviation (std) of the gaussian noise. Must be broadcastable with signal if a Tensor. If a 2-tuple of float (min, max), the resulted std will vary along dimensions dims and follows a uniform distribution in this range.

  • noise_only (bool) – If True, return noise rather than noisy signal. Default: False.

  • dims (int | Sequence[int]) – Dimensions with varied standard deviation. See description for sigma. Default: 0.

Returns:

Noisy signal.

Return type:

Tensor