Denoise PPG using a user-supplied autoencoder reconstruction model — denoise_gazepoint_ppg

Applies a user-supplied reconstruction function or model to fixed-length PPG windows. No pretrained neural network is bundled. This function is an interoperability bridge for validated user-supplied autoencoders.

Usage

denoise_gazepoint_ppg_autoencoder(
  dat,
  ppg_col = "HRP",
  time_col = NULL,
  group_cols = NULL,
  model = NULL,
  window_samples = 128,
  output_col = NULL,
  overwrite = FALSE
)

Arguments

dat: A data frame.
ppg_col: Numeric PPG/pulse column.
time_col: Optional time column.
group_cols: Optional grouping columns.
model: A user-supplied function or model. If NULL, the original signal is copied and status records that no model was supplied.
window_samples: Window length in samples.
output_col: Optional output column.
overwrite: Logical. If FALSE, existing output column is protected.

Value

A data frame with reconstructed signal and denoising attributes.