Estimation of the cooking progress of deep-fried foods using machine learning and sound analysis

Abstract

In order to promote the automation of cooking equipment, automatic estimation of the cooking progress is expected to take over the role of elapsed-time counter. For this purpose, we conducted a set of experiments in which the cooking progress along time is estimated using acoustic feature analysis and machine learning. The results suggests that precise estimation of the cooking progress is feasible by using the acoustic feature and machine learning. Moreover, it was revealed what feature was necessary to judge the cooking progress of deep-fried foods, and the required condition for temporal analysis was clarified.