1. Problem

Context plays a crucial role in understanding emotions in people. Emotions are multi-layered and are significantly influenced by the individual situation, environment, relationships, and past experiences. Traditionally, emotions and their context have been recorded in diaries. This approach allowed for personal reflection and a better understanding of one's emotions and connections to everyday life. Today, these diaries are increasingly being transferred to a digital format targeting the automatic recognition of emotions and their relation to specific context. Various classifiers are already available for emotion recognition using biosignals and labels captured by smartwatches . However, despite the great potential, contextual information is neither captured nor leveraged for improving the classifiers systematically.

2. Goal

The goal of this bachelor thesis is to solve the aforementioned problem by enhancing an existing smartwatch app for emotion labeling to support context labeling. The app should record the data from the accelerometer of the smartwatch and the user should be supported to label this context data with a simple and user-friendly interface. The app should be developed for Samsung smartwatches using Google Wear OS. As part of the this a pilot data collection with test persons should be carried out. An existing classifier for emotion classification can be leveraged to analyze the potential of adding contextual information to the classification problem.

3. Requirements

• Strong time management skills
• First Java/Kotlin experiences or at least an interest in learning the technology
• Interest in user interface design for smartwatches
• Interest in biosignals and their processing