Topic: Differentiating Eating Behaviors Using Video-Based Mouth Movement Detection

Topic: Differentiating Eating Behaviors Using Video-Based Mouth Movement Detection

Personal details

Title Differentiating Eating Behaviors Using Video-Based Mouth Movement Detection
Description
Project Overview:
Are you passionate about applying computer vision techniques to impactful health and nutrition research? This master’s thesis project builds on cutting-edge work in video-based mouth movement detection, initially developed for speech recognition. The aim is to extend this approach to identify and differentiate various eating behaviors, such as distinguishing between eating and drinking, classifying food textures (e.g., hard vs. soft), and estimating food or liquid consumption.
By analyzing video data, this project aims to develop an advanced system for monitoring eating habits in both clinical and real-world settings. The system could revolutionize healthcare, dietary monitoring, and behavioral research by offering a non-invasive, automated method to track food intake.
 
Opportunities:
  • Potential to publish research findings in leading computer vision and healthcare conferences.
  • Engage with an interdisciplinary research team, receiving guidance from experts in machine learning, computer vision, and healthcare.
  • Hands-on experience applying AI to real-world healthcare challenges.
Supervision:
You will receive regular support from a team of experts. The project offers access to collaborative research resources and practical experience working with Raspberry Pi-based systems.
 
How to Apply:
Interested candidates should submit their CV, a brief statement of interest, and details of any relevant research experience to rebecca.diekmann@uni-oldenburg.de  and martin.bleichner@uni-oldenburg.de. We look forward to hearing from you!

 

Home institution Department of Psychology
Associated institutions
Type of work practical / application-focused
Type of thesis Master's degree
Author Dr. Martin Georg Bleichner
Status available
Problem statement
Key Objectives:
  1. Extension of Video-Based Mouth Movement Models: Adapt existing models (originally used for speech detection) to classify mouth movements related to eating and drinking.
  2. Differentiation of Eating and Drinking: Develop a computer vision algorithm capable of distinguishing between drinking and various eating behaviors through mouth movement analysis.
  3. Food Texture Classification: Expand the model to classify food textures (e.g., hard vs. soft foods) by identifying patterns in mouth movement.
  4. Quantification of Consumption: Explore techniques to estimate the amount of food or liquid consumed based on visual data, focusing on aspects like mouth size and chewing patterns.
  5. Integration with Hardware (Raspberry Pi): Implement and test the models on a Raspberry Pi (Zero) to create a portable, real-time system for tracking eating behaviors.
Requirement
Candidate Profile:
  • Ideal Background: Students with an interest in computer vision, machine learning, and healthcare technologies.
  • Desirable Skills (Nice to Have or Willing to Learn):
    • Experience with Python and computer vision libraries (e.g., OpenCV) is beneficial but not required, as you'll have opportunities to learn these skills.
    • Familiarity with deep learning frameworks like TensorFlow or PyTorch is a plus but can be developed during the project.
    • Knowledge of neural networks (e.g., ResNet and its variations) is advantageous but not essential.
    • Basic knowledge or willingness to work with Raspberry Pi for deploying and testing models is appreciated.
Created 16/10/24

Study data

Departments
Degree programmes
  • Master's Programme Neuroscience
  • Master's Programme Neurocognitive Psychology
  • Master's Programme Engineering Physics
  • Master's Programme Engineering of Socio-Technical Systems
Assigned courses
Contact person