Table 4.

Comparison of previous studies on food intake episodes detection.

Study	Modalities	Method	Accuracy, %
[32]	Acoustic signal	Correlation matching	85
[33]	Food image and speech recording	Support vector machine classification	90.6
[34]	Electroglottograph	Artificial neural network	86.6
[35]	Piezoelectricity	Time and amplitude thresholding	86
[36]	Accelerometer and gyroscope	Decision tree classifier	85.5
[37]	Chewing sound	(1) Deep Boltzmann and (2) Machine with deep neural network classifier	77
[38]	Piezoelectricity	Convolutional neural network	91.9
[39]	Acceleration and orientation velocity	Convolutional-recurrent neural network	82.5-96.4