In this work the application of an additional algorithm for the identification and validation of signals generated by the nucleation of bubbles in calibrated superheated droplet detectors is presented; more specifically, events having low amplitudes (< 2 mV) in the order of the noise level are here investigated. After data filtering, a peak finding algorithm for the purpose of counting all the events has been implemented; the performance of the algorithm demonstrates a factor 40 reduction in the noise level. Such technique permits to identify and count events also in considerable large data sets, like the ones obtained during calibrations. Experimental results are extrapolated from the analysis of detector calibration data set. These results show an increase in the counting effectiveness of events (2%–5%) having low amplitude and originated in detectors characterized by a low size droplet distribution.