Remarks on neural network-based tea aroma recognition with a mass-sensitive chemical sensor using plasma-organic-polymer-film-coated quartz crystal resonators

This study investigates a method for recognition of tea aroma using a multilayer neural network. A gas-sensing system utilising an array of quartz crystal resonators coated with plasma-organic-polymer-films is designed. Three teas, oolong, jasmine and puerh, are used as the sources of sample gas. Wavelet transform frequency analysis is applied to the signals from the gas-sensing system, and the features of the sensor response are extracted. These features are used as input vectors to a neural network. The neural network achieves an average recognition rate of 96.7% for the three tea samples. Our experimental results demonstrate the effectiveness of the gas-sensing system and neural network in this application.