Nourich

From Knoesis wiki
Jump to: navigation, search

Motivation and Background

Obesity, being one of the diseases with major comorbidities, is on the rise worldwide. In the United States, 36% of the adult population are obese [1]. Obesity can be caused by genetic and behavioral factors which include irregular eating patterns, lack of physical activity and the side-effect of medication usage [2]. According to The Dietary Guidelines of Americans 2010, focus on reducing excess calorie consumption and making an informed decision about food choices and physical activity can help attain a healthier weight and reduce the risk of chronic illness. While the weight trend can move towards a healthier range by controlled calorie consumption with essential nutrition intake, people/patients are unable to determine a healthy regimen for various reasons. Specifically, they are unaware of calorie content in the food they intake because it is not easy to determine it, and even if they do, keeping track of their food pattern and cumulative consumption is difficult.

kHealth kit for Bariatrics

Nourich

Monitoring an individual's diet and cumulative calorie intake through food images and recommending meals can help them in making informed decisions about their meals. Also, tracking and assessing their food patterns and weight trends can help them maintain a healthier weight in the longer run. A system can be built that is trained to recognize food images collected from open sources such as Instagram, Google images, Pinterest, Getty image, etc. Once recognized, the volume can be estimated based on user input (automatically, in future) and nutrition information can be obtained using comprehensive knowledge bases. AI techniques support meal recommendations specific to user preferences and context.

Applications

Android App

Nourich has an Android app version which is trained on food images and can recognize food labels from images. It uses one of the world's largest nutrition knowledge base, Edamame. Right now, the volume is input by the user and will be automatically estimated later(future work). The system architecture is below.

Android app architecture

Chatbot

Nourich also has a chatbot version, a voice assistant that includes leveraging two knowledge graphs (domain-specific and personalized user-centric) to (i) provide long-term tracking of food log, diet adherence, and daily calorie intake in a ubiquitous manner and (ii) identify the diet habits of the users through images and incorporate them to profile the users to deliver personalized and effective dieting guidelines.

The design and architechture of Nourich can be found below. It is capable of combining different, yet complementary modules and technologies to support various functionalities as follows.

    • 1) NOURICH will first request the user for a set of personal information. This initial profile setup is necessary for the voice assistant to get to know the user better and to establish a baseline profile to populate the user knowledge graph.
    • 2) The user is monitored for an indefinite duration based on the his/her engagement. NOURICH will pop up a daily chat notification to interactively converse with the user for food logging. NOURICH allows dynamic engagement and continuous monitoring of User/Patient-Generated Health Data (PGHD) to (i) track and illustrate diet changes over time, (ii) understand the effects of specific diet regime, (iii) understand user’s diet adherence, and (iv) devise personalized food recommendations that revolve around the user’s food and diet preference. The longer NOURICH monitors the user, the better it gets to know its user.
    • 3) User is able to ask food and diet-related questions which are answered by NOURICH using the knowledge available in various food ontologies. This allows the user to be well-informed of the food and calories that he/she is consuming.
    • 4) The system augments personalized health information using the contextualized individual user knowledge graph to support various stages of augmented health management strategies.


Android app architecture

Contact

References

  1. “Adult Obesity Facts | Overweight & Obesity | CDC.” Centers for Disease Control and Prevention, Centers for Disease Control and Prevention, www.cdc.gov/obesity/data/adult.html.
  2. “Adult Obesity Causes & Consequences | Overweight & Obesity | CDC.” Centers for Disease Control and Prevention, Centers for Disease Control and Prevention, www.cdc.gov/obesity/adult/causes.html.