The determination of peaks in biological waveforms.

A method to determine the primary peaks of a waveform, in the presence of noise and a drifting baseline, is proposed. The method is based on a definition of a peak in terms of its height relative to adjacent valleys. Two algorithms to determine the set of primary peaks are presented. Some theoretical results, which ensure the stability of the algorithms, are given. The approach is illustrated by the analysis of chromatogram peaks.