Non-invasive monitoring of eating behavior using spectrogram analysis in a wearable necklace

Abstract

Food intake levels, hydration, chewing and swallowing rate, and dietary choices are all factors known to impact one's health. This paper presents a novel wearable system in the form of a necklace, which aggregates data from an embedded piezoelectric sensor capable of detecting skin motion in the lower trachea during ingestion. We propose an algorithm based on spectrogram analysis of piezoelectric sensor signals to accurately distinguish between food types such as liquid and solid, hot and cold drinks and hard and soft foods. The necklace transmits data to a smartphone, which performs the processing of the signals, classifies the food type, and provides visual feedback to the user to assist the user in monitoring their eating habits over time. Experimental results demonstrate high classification accuracy of the proposed method, and validate the use of a spectrogram in extracting key features representative of the unique swallow patterns of various foods.

Food Intake Detection and Classification Using a Necklace-Type Piezoelectric Wearable Sensor System

“…Piezoelectric sensors embedded in stretchable necklaces are available in two different forms: stretchable sports bands and pendants [22]. Unlike [7], [13], [19], [23], we have se- lected a stretchable sports band design. This band can be adjusted to the neck diameter of the user, so people of different physiques can wear it without any extra effort.…”

“…The sports band maintains high priority in algorithm evolution, data collection and clinical environments. The second necklace design, the pendant [13], [19], [23], has the disadvantage of leaving its target position during motions such as walking, running or vigorous head-turning. Therefore, a stretchable sports band necklace has been used to embed a piezoelectric sensor in our envisioned system, as shown in Fig.…”

“…Our designed system not only helps detecting food intake but also recognizes food categories in real-life conditions. We demonstrate that carefully designed raw data Copyright c 2018 The Institute of Electronics, Information and Communication Engineers collection in the time domain, along with an optimal feature selection, enables our method to outperform the state-ofthe-art piezoelectric sensor-based method [13], [19] that employs spectrogram features for food recognition. We have found that, a prior chewing sequence combined with a successive swallowing event, forming a temporal pattern of 70 samples (3.5 seconds) with swallowing located towards the end of the frame, generates the most effective features for classification.…”

Recognition of Nutrition Intake Using Time-Frequency Decomposition in a Wearable Necklace Using a Piezoelectric Sensor

“…Current systems either have low accuracy in detecting swallows and distinguishing food types, or must be uncomfortably worn around the neck, which renders continuous use impractical [8]. In this paper, we focus on a piezoelectric-based design of a necklace that is not worn tightly around the neck, but rather hangs loosely and falls more naturally right above the sternum [24], [5], [20], [21], [23].…”

Embedded design consideration for the implementation of food intake monitoring system:validated using LabVIEW software

“…The whole hardware design is bulky and the device is minimized to thehospital environment. Detection of food intake have been explored by following several other methods [1], [2], [5]introduced food intake detection by sensing arm gestures using accelerometers and gyroscope sensors. Food intake is monitored by identifying the arm gestures associated with thehandling of cups, spoon, and plates.…”