AI-enabled smart glasses for continuous monitoring of food and water intake and personalized nutrition support

Abstract

A wearable eyewear device for real-time monitoring of food and water intake, comprising: • a spectacle frame integrated with a processor; • at least one high-resolution camera configured to capture images of food; • a depth sensor to estimate the portion volume; • a food recognition module that uses AI to identify the type of food and its nutritional content; • a barcode / OCR module for reading data on packaged food; • a bone conduction speaker and microphone for audio feedback and voice control; • a transparent HUD for visual warnings; • a sensor array with gyroscope, accelerometer and ambient light sensor; • an alert module for warnings based on the user's dietary limits; • a water intake detection module that analyses drinking gestures; • a data storage / transfer system that can be synchronized with external devices.

Description

SCOPE OF THE INVENTION

[0001] The present invention relates to a wearable health device in the form of AI-integrated smart glasses designed for continuous real-time monitoring of food and water intake, nutritional analysis, calorie estimation, and personalized nutrition advice. This system passively tracks dietary habits, provides real-time alerts, and offers voice-based interaction, making it particularly suitable for individuals whose health conditions require dietary regulation, such as diabetes or cardiovascular disease, as well as for visually impaired users who require accessible nutritional assistance. BACKGROUND

[0002] Conventional dietary tracking systems rely heavily on manual data entry into mobile applications or fitness wearables with limited functionality. These methods are not designed for hands-free, real-time monitoring and are often inaccessible to users with visual impairments. Existing wearables lack AI-driven insights tailored to specific health conditions such as diabetes, kidney problems, or hypertension, and do not seamlessly integrate dietary monitoring with accessible features. There is a critical need for a non-intrusive, AI-enabled wearable that passively monitors food intake, provides real-time nutritional recommendations, and assists visually impaired users through voice-based interaction.The present invention addresses these challenges by providing an automated, accessible, and personalized solution for nutrition management. SUMMARY OF THE INVENTION

[0003] The AI-enabled smart glasses are eyeglass-like wearable devices equipped with cameras, sensors, AI processors, barcode readers, and audiovisual modules for real-time monitoring and advice on food and water intake. The system captures images of food and beverages, identifies items and portion sizes using AI-based image processing, and estimates nutritional and calorie values ​​using a pre-trained food database. It provides visual and voice-based alerts when food intake exceeds preset medical limits, reads nutrition labels for visually impaired users using OCR, and supports voice-based customization of dietary preferences.Data is logged and synchronized with a companion mobile or desktop application for detailed analytics, allowing users, caregivers, or healthcare professionals to track dietary trends and ensure adherence to health goals. DETAILED DESCRIPTION

[0004] The AI-enabled smart glasses consist of several key components integrated into a comfortable, aesthetically pleasing frame, ensuring continuous data collection and user interaction: 1. Key components: • High-resolution dual cameras: Embedded on the left and right temples near the hinges, which capture images of food, detect drinking motion, and estimate portion size. • IR depth sensor: Located near the camera modules, it helps estimate portion volume. • AI Main Processor (SoC) & NPU: Housed in the right temporal arm, which performs food detection, AI models and audio processing with accelerated inference. • Gyroscope and accelerometer: Located in both temporal arms, they detect head tilt and movement to track water intake. • Bone conduction speaker: Located on the temporal arm near the ear, it provides audio feedback and alerts without obstructing ambient noise. • MEMS microphones: Embedded in the lens corners or bottom frame, which receive voice commands and ambient noise. • Transparent HUD display: Projected into the right lens and shows visual warnings, icons and recording summaries. • Barcode / OCR scanner module: Embedded in the camera area of ​​the nose bridge, reads barcodes, QR codes and nutrition labels. • Ambient light sensor: Located near the lens frame corner, adjusts camera exposure and assists OCR in low-light conditions. • Battery (1000-1200mAh): Embedded in both arms, which supplies power to all components. • USB-C charging port: Under the right temple arm, retractable or magnetic, for charging devices. • NFC / QR reader module: Located near the camera unit, enables quick scanning of food packaging. 2. Software and AI architecture: • Food recognition module: Uses pre-trained datasets (e.g., Food-101, UEC-Food256) to classify food types (cooked, raw, packaged) and applies deep sensing for quantity estimation. • Calorie and Nutrition Estimation Module: Matches foods to a nutritional database and estimates calories, sugars, carbohydrates, fats, proteins and sodium, taking into account portion volume. • Health-based filter: Compares nutritional data with user health profiles (e.g., diabetes, high blood pressure) and triggers alerts when thresholds are exceeded. • OCR and Visual Aid Module: Reads barcodes, QR codes and labels and uses text-to-speech to assist visually impaired users. • Water Intake Detection: Analyzes head tilt and drinking posture using motion sensors and AI image recognition to log fluid intake. • Alarm system: Displays red / green indicators on the HUD and provides audio warnings (e.g., “High salinity detected”). • Voice interaction: Supports NLP-based voice commands (e.g., “What did I eat today?” or “Read this label”). • Data recording and synchronization: Stores data locally and synchronizes it with a companion app for dashboards, charts, and reports. 3. Operating process: • The cameras take pictures of food or drinks while the user prepares to eat or drink. • The AI ​​processor identifies items, estimates portion sizes using depth sensors, and calculates nutritional values. • The health filter checks food intake based on medical limits and triggers HUD or audio warnings if these limits are exceeded. • The OCR module scans labels and reads nutritional information aloud for visually impaired users. • Motion sensors detect drinking gestures and record water intake. • Users give voice commands to adjust settings or check food intake history. • The data is synchronized with the companion app and can be reviewed by users, caregivers or doctors. 4. Companion Application Ecosystem: • Offers food diaries, nutrient limit settings, physician input integration, and compatibility with smartwatches, fitness bands, and electronic health records (EHRs). • Supports cloud synchronization and real-time alerts for caregivers for remote monitoring.

Claims

[1] A wearable eyeglass device for real-time monitoring of food and water intake, comprising: • a spectacle frame integrated with a processor; • at least one high-resolution camera configured to capture images of food; • a depth sensor to estimate the portion volume; • a food recognition module that uses AI to identify the type of food and its nutritional content; • a barcode / OCR module for reading data on packaged food; • a bone conduction speaker and microphone for audio feedback and voice control; • a transparent HUD for visual warnings; • a sensor array with gyroscope, accelerometer and ambient light sensor; • an alert module for warnings based on the user's dietary limits; • a water intake detection module that analyses drinking gestures; • a data storage / transfer system that can be synchronized with external devices. [2] The apparatus of claim 1, wherein the food recognition module is configured to adapt through voice input-based training. [3] The device of claim 1, further comprising a visual assistance module that reads nutrition labels aloud to assist visually impaired users. [4] The device of claim 1, wherein the AI ​​engine calculates calorie / nutrient data based on portion estimates using camera and IR sensor fusion. [5] The device of claim 1, wherein the warning system operates via a semi-transparent HUD and bone-conducting audio signals. [6] The device of claim 1, wherein the processor stores historical food / water intake data for trend analysis and medical reports. [7] The device of claim 1, wherein the device communicates wirelessly with a companion mobile application for real-time monitoring. [8] The apparatus of claim 1, wherein water uptake is estimated using motion capture and visual object recognition algorithms.