Standardize windows of signal data — standardize

Standardizes windowed objects by applying various transformations to each window. This function converts each egm object in a windowed list to a standardized data frame with uniform properties, facilitating comparison and analysis.

Usage

standardize_windows(
  x,
  standardization_method = c("time_normalize"),
  target_samples = 500,
  target_ms = NULL,
  interpolation_method = c("linear", "spline", "step"),
  align_feature = NULL,
  preserve_amplitude = TRUE,
  preserve_class = FALSE,
  ...
)

Arguments

x: A windowed object to standardize
standardization_method: A character string specifying the standardization method. Currently supported: "time_normalize".
target_samples: The desired number of samples for each standardized window. Default is 500 samples. This parameter takes precedence if both target_samples and target_ms are provided.
target_ms: Alternative specification in milliseconds. If provided and target_samples is NULL, the function will convert this to samples based on the signal's sampling frequency.
interpolation_method: The method used for interpolation when resampling. Options are "linear" (default), "spline", or "step".
align_feature: Feature to align windows around, either a character string matching an annotation type or a list of criteria for finding a specific annotation. Default is NULL (no alignment).
preserve_amplitude: Logical. If TRUE (default), maintains original amplitude range after resampling.
preserve_class: Logical. If TRUE, returns a windowed object with standardized data frames. If FALSE (default), returns a plain list of data frames.
...: Additional arguments passed to specific standardization methods.

Value

If preserve_class=TRUE, a windowed object containing standardized data frames. If preserve_class=FALSE, a plain list of standardized data frames.

Details

Currently supported standardization methods:

time_normalize - Resamples each window to a standard length by either dilating or contracting the signal. The result is a signal with a consistent number of samples regardless of the original window duration.

Additional options:

align_feature - If provided, windows will be aligned to center around this feature (e.g., a specific annotation type like "N" for R-peak). Can be a character string matching an annotation type or a list of criteria for annotation matching.
preserve_amplitude - If TRUE (default), maintains the original amplitude range after resampling. If FALSE, the amplitudes may change due to interpolation.

Examples

if (FALSE) { # \dontrun{
# Read in ECG data
ecg <- read_wfdb("ecg", test_path(), "ecgpuwave")

# Create windows based on sinus rhythm
windows <- window_signal(
  ecg,
  method = "rhythm",
  rhythm_type = "sinus",
  onset_criteria = list(type = "(", number = 0),
  offset_criteria = list(type = ")", number = 2),
  reference_criteria = list(type = "N")
)

# Standardize windows to exactly 500 samples
std_windows <- standardize_windows(
  windows,
  method = "time_normalize",
  target_samples = 500
)

# Alternatively, standardize to 500 milliseconds (depends on sampling frequency)
std_windows_ms <- standardize_windows(
  windows,
  method = "time_normalize",
  target_ms = 500
)

# Standardize windows with QRS alignment
aligned_windows <- standardize_windows(
  windows,
  method = "time_normalize",
  target_samples = 500,
  align_feature = "N"  # Align on QRS complexes
)
} # }