denspp.offline.metric.snr#

Module Contents#

Functions#

calculate_snr

Calculating the signal-to-noise ratio [dB] of the input signal compared to mean waveform

calculate_snr_tensor

Calculating the Signal-to-Noise (SNR) ratio of the input data Args: data: Tensor with raw data / frame mean: Tensor with class-specific mean data / frame Return: Tensor with SNR value

calculate_dsnr_tensor

Calculation of metric different Signal-to-Noise ratio (SNR) between defined input and predicted to reference waveform Args: data: Tensor array with input waveform pred: Tensor array with predicted waveform from model mean: Tensor array with real mean waveform from dataset Return: Tensor with differential Signal-to-Noise ratio (SNR) of applied waveforms

calculate_snr_cluster

Calculating the cluster-specific Signal-to-Noise Ratio (SNR) for all frames

API#

denspp.offline.metric.snr.calculate_snr(data: numpy.ndarray, mean: numpy.ndarray) → numpy.ndarray[source]#

Calculating the signal-to-noise ratio [dB] of the input signal compared to mean waveform

Parameters:

  • data – Numpy array with all spike waveforms (raw data)

  • mean – Numpy array with mean waveform of corresponding spike frame cluster

Returns:

Numpy array with SNR of all spike waveforms

denspp.offline.metric.snr.calculate_snr_tensor(data: torch.Tensor, mean: torch.Tensor) → torch.Tensor[source]#

Calculating the Signal-to-Noise (SNR) ratio of the input data Args: data: Tensor with raw data / frame mean: Tensor with class-specific mean data / frame Return: Tensor with SNR value

denspp.offline.metric.snr.calculate_dsnr_tensor(data: torch.Tensor, pred: torch.Tensor, mean: torch.Tensor) → torch.Tensor[source]#

Calculation of metric different Signal-to-Noise ratio (SNR) between defined input and predicted to reference waveform Args: data: Tensor array with input waveform pred: Tensor array with predicted waveform from model mean: Tensor array with real mean waveform from dataset Return: Tensor with differential Signal-to-Noise ratio (SNR) of applied waveforms

denspp.offline.metric.snr.calculate_snr_cluster(frames_in: numpy.ndarray, frames_cl: numpy.ndarray, frames_mean: numpy.ndarray) → numpy.ndarray[source]#

Calculating the cluster-specific Signal-to-Noise Ratio (SNR) for all frames

Parameters:

  • frames_in – Numpy array with spike frames

  • frames_cl – Numpy array with cluster label to each spike frame

  • frames_mean – Numpy array with mean waveforms of cluster

Returns:

Numpy array with SNR value for each sample for {min, mean, max}