What if Your Home Wi-Fi Could See You Eat?

WiEat: A Device-Free Auditor

By repurposing the ubiquitous 5 GHz Wi-Fi signals already saturating our homes, the system can track the exact mechanics of a meal with startling precision. It turns a standard smartphone or laptop into a sophisticated health auditor.

This matters because human memory is notoriously fickle. Most dietary data relies on self-reporting, which is plagued by subjective bias and simple forgetfulness. By automating the "log," WiEat offers a way to monitor metabolic health, eating speed, and habits without the user ever having to press a button or wear a wristband.

The Core Methodology

The study observed 20 participants over a total of 1600 minutes of eating data. It utilized Channel State Information (CSI) to map movements.

When a person moves their hand from a plate to their mouth, they disturb the Wi-Fi signal’s path. The system is so sensitive that it doesn't just see the arm moving; it captures the "micro-scale" vibrations of facial muscles during mastication and deglutition (chewing and swallowing).

Detection Performance

The results suggest our wireless routers are surprisingly good detectives. The system achieved a 100% detection rate for eating activities** across both smartphone and IoT configurations.

Using a "Soft Decision" probability strategy, it identified specific utensils with surgical accuracy:

• 95.29% for spoons
• 93.07% for forks
• 94.28% for bare hands

Measuring the Rhythm of Eating

Even the rhythm of the meal was quantified. The system estimated chewing frequency with a percentage error of just:

• 9% for IoT devices
• 13% for smartphones

To achieve this, the team filtered for minute motions within the 0.8 Hz–3 Hz range, effectively isolating the pulse of a jawline from the background noise of a living room.

The System's Blind Spots

However, the "invisible eye" has its blind spots. The researchers noted key limitations:

• Range & Interference: While robust up to 120 cm, its performance degrades if another person is moving within 2 meters, making crowded restaurants a challenge.
• Gesture Mimicry: It struggles with "mimic" motions. The study specifically excluded smoking because the hand-to-mouth gesture is too similar to eating.