Exploring ML use cases
ML is rapidly evolving to more use cases such as fraud detection, personalized targeted marketing, and self-driving cars, to name a few. Currently, the most popular applications for ML in the embedded space can be grouped into three areas: vibration, vision, and voice. These categories were briefly discussed in Chapter 1, Selecting the Right Hardware, with the current chapter focusing on running examples in each of these areas. Vibration typically is the most straightforward and is therefore a good starting point to begin exploring. After familiarizing ourselves with the vibration example, we provide more details on the construction of the ML software stack in the vision and voice examples that follow. The three use cases are presented here:
- Anomaly detection using the Cortex-M55
- Image classification using the Cortex-M55 and Ethos-U55
- Micro speech using the Cortex-M55 and Ethos-U55
Anomaly detection – vibration
To replicate the...