Master Thesis Assignment: Head Gesture Recognition using the OpenEarable Platform
Open Earable
Introduction
Wearable devices equipped with various sensors have the potential to revolutionize various fields, including healthcare, fitness, well-being and assistive technologies. The OpenEarable platform offers an open-source framework, https://github.com/OpenEarable/open-earable, for developing and deploying applications on ear-worn devices. Head gesture recognition has the potential to revolutionize human-computer interaction, enabling intuitive and hands-free control in various domains. This thesis project aims to explore the capabilities of the OpenEarable platform for head gesture recognition using its embedded sensors.
Objectives
- Set up the OpenEarable hardware platform, including sensor configuration and software installation.
- Collect sensor data from the OpenEarable platform during various head movements (e.g., nodding, shaking, tilting).
- Develop and implement a machine learning algorithm on the collected data to classify different head gestures.
- Evaluate the performance of the head gesture recognition system in terms of accuracy, precision, recall, and other relevant metrics.
Project Description
You will be tasked with:
- Literature Review: Conduct a comprehensive review of relevant literature on the OpenEarable platform, wearable sensor technology.
- Platform familiarization: Gain in-depth knowledge of the OpenEarable platform through tutorials, documentation, and community resources.
- Hardware Setup: Assemble the OpenEarable hardware platform according to the provided instructions.Configure and calibrate the onboard sensors, including the inertial measurement unit (IMU). Install and configure necessary software libraries and development tools.
- Data Collection: Design a set of experiments to collect sensor data during various head movements. Implement data collection software on the OpenEarable platform to record sensor data with timestamps.
- Head Gesture Classification: Preprocess the collected sensor data for machine learning. Explore and implement various machine learning algorithms (e.g., decision trees, random forests, support vector machines) for head gesture classification. Train and evaluate the chosen machine learning model on the collected data. Evaluate the performance in terms of accuracy, precision, recall, and other relevant metrics.
- Report Writing: Document all the stages of the project, including the design process, data analysis, system testing, and evaluation results.
Pre-Requisites
Prospective students should have a background in Computer Science or Embedded Systems or Electrical Engineering, or a related field. Knowledge of hardware integration, data analysis and classification will be beneficial.
Are you ready to take up this challenge and contribute to emerging field of hearables/earables?
Contact
Özlem Durmaz, Associate Professor, Pervasive Systems (ozlem.durmaz@utwente.nl)