Extract heartbeat candidates from Gazepoint pulse using k-means — extract_gazepoint_beats

Uses k-means clustering on the raw pulse waveform to classify likely heartbeat regions and then selects local extrema as beat candidates. This is a Gazepoint Biometrics-oriented fallback for difficult pulse waveforms, not an ECG-equivalent R-peak detector.

Usage

extract_gazepoint_beats_kmeans(
  dat,
  pulse_col = "HRP",
  time_col = "CNT",
  group_cols = NULL,
  k = 2,
  peak_polarity = c("positive", "negative"),
  min_distance_s = 0.3,
  sampling_rate = NULL,
  seed = NULL
)

Arguments

dat: A data frame.
pulse_col: Numeric pulse/PPG column.
time_col: Numeric time column.
group_cols: Optional grouping columns.
k: Number of k-means clusters.
peak_polarity: "positive" or "negative".
min_distance_s: Minimum time between selected beats.
sampling_rate: Optional sampling rate in Hz.
seed: Optional random seed.

Value

A list with overview, beat_table, interval_table, timeseries, and settings.