Signal Processing for Interactive Systems

Extracting meaningful features from your data is crucial to building small and reliable machine learning models.

On Edge Impulse this is done through processing blocks: a number of processing blocks for common sensor data (such as vibration and audio).

• Raw Data

• Flatten

• Image

• Spectral features

• Spectrogram

• Audio MFE

• Audio MFCC

• Audio Syntiant

• IMU Syntiant

• HR/HRV features

The source code of these blocks is available in the Edge Impulse processing blocks GitHub repository.

Custom processing blocks

If you have a very specific sensor, want to apply custom filters, or are implementing the latest research in digital signal processing, follow the Edge Impulse tutorial on Building custom processing blocks.

To pass SPIS, you will program your own custom DSP block, fine-tune and apply it to a specific problem and deliver eihter as a notebook or as an Edge Impulse block. The application could be either your own problem (classical mini-project) or a hackathon (TBA). The total duration of the work is about 30 hours (2 ECTS).