select_signal_TFpeaks¶
- select_signal_TFpeaks(tpeak_properties, plot_on)¶
SELECT_SIGNAL_TFPEAKS Select TF-peak signals via k-means on duration and bandwidth dimensions
- Usage:
[signal_idx, fig_compare] = select_signal_TFpeaks(tpeak_properties, plot_on)
- Inputs:
tpeak_properties : struct - time-peak properties (proms, durations, central_frequencies, bandwidths) – required plot_on : logical - show diagnostic histograms (default: false)
- Outputs:
signal_idx : Nx1 logical - true where the peak is labeled as signal fig_compare : handle - figure handle (empty when plot_on = false)
Notes
Runs k-means on each of 5 derived features (log-prominence, log-volume, duration, central frequency, bandwidth) with sensible class priors on duration ([0, 0.3, 1]) and bandwidth ([0.5, 2, 3.5]). The signal label is the conjunction of the right-most cluster on duration AND on bandwidth.
See also: TF_peak_detection, TF_peak_selection, kmeans
∿∿∿ Prerau Laboratory MATLAB Codebase · sleepEEG.org ∿∿∿