extract.fp.minmax: Extract Time and Strength Independent Fingerprints

In LiKao/VoiceExperiment: Analysis of Vocal Responses to Psychological Experiments

Description

Extract Time and Strength Independent Fingerprints

Usage

extract.fp.minmax(fs)

Arguments

fs	The feature time series from which the fingerprints should be extracted

Details

This function extracts fingerprints, which are somewhat invariant to stretching in the time domain while keeping the order of features intact. Furthermore these fingerprints are completely invariant to a scaling or offset of the complete feature series.

For each feature within the feature vectors, the times of maximum and minimum value are extracted. Since no values of the actual features are used, scaling and offset invariance is achieved. Furthermore, the positions of minima and maxima are scaled relative to the total length of the series. Therefore, these fingerprints are robust to stretching of the complete time series. Local stretching of the feature series are translated to movements in euclidean space.

LiKao/VoiceExperiment documentation built on May 7, 2019, 1:19 p.m.