The gesture content pack will allow Imagimob AI customers to quickly and easily prototype, develop and test gesture applications based on the Acconeer radar sensor A1. By collecting more data, the current AI-model can be improved, and new gestures can be added.
The content pack includes:
- A trained AI-model (in h5-format) that recognises three defined gestures
- Annotated dataset with metadata (video)
- Pre-processing that are necessary to generate highly efficient edge-models
- Capture server – a software that combines with an Android app for collecting more data. Or a local version that uses a webcame for a stand-alone solution.
- Python-script that allows for live testing on a PC
Acconeer development kits can be purchased from Acconeer's distributors, eg Digi-Key.
The gesture content pack is available from February 1, 2021.