Signaligner Pro: A Tool to Explore and Annotate Multi-day Raw Accelerometer Data.

Area of Science:

• Human-Computer Interaction
• Wearable Technology
• Machine Learning

Background:

• Accurate human activity recognition (HAR) using wearable accelerometers is crucial for novel human-computer interfaces.
• Machine learning HAR algorithms often require extensive, precisely annotated multi-day datasets, which are difficult to create.
• Existing tools for visualizing and annotating high-sampling-rate accelerometer data are limited.

Purpose of the Study:

• To introduce Signaligner Pro, an interactive software tool designed for efficient exploration and annotation of multi-day, high-sampling-rate raw accelerometer data.
• To address the scarcity of user-friendly tools for generating annotated datasets for HAR research.
• To facilitate the creation of improved annotated datasets by researchers.

Main Methods:

• Development of Signaligner Pro, an interactive visualization and annotation tool.
• Implementation of features for exploring high-sampling-rate raw accelerometer data.
• Integration of functionality to visualize and directly modify existing time-stamped annotations from algorithms and human annotators.

Main Results:

• Signaligner Pro enables convenient visualization and annotation of multi-day accelerometer data.
• The tool supports the direct modification of annotations, allowing for dataset improvement.
• Demonstrated utility in generating accurate activity annotations for HAR.

Conclusions:

• Signaligner Pro effectively supports researchers in creating high-quality annotated datasets for human activity recognition.
• The tool streamlines the process of data preparation for machine learning models.
• Facilitates advancements in the field of wearable sensor-based activity recognition.