- datasets placeholder for the Hand Posture ToF datasets.
- docs contains all readmes and documentation specific to the hand posture recognition use case.
- tf contains Tensorflow based tools to train, evaluate, benchmark and deploy your model on your STM32 target.
- config_file_examples contains example configuration files for different operation modes.
The operation_mode top-level attribute specifies the operations or the service you want to execute.
The different values of the operation_mode attribute and the corresponding operations are described in the table below.
All .yaml configuration examples are located in the config_file_examples folder.
| operation_mode attribute | Operations |
|---|---|
training |
Train a model from the model zoo or your own model |
evaluation |
Evaluate the accuracy of a float model on a test or validation dataset |
benchmarking |
Benchmark a float model on an STM32 board |
deployment |
Deploy a model on an STM32 board |
You can refer to the README links below that provide typical examples of operation modes and tutorials on specific services:
The hand posture use case is based on the ST multi-zone Time-of-Flight sensors: VL53L5CX and VL53L8CX. The goal of this use case is to recognize static hand postures such as a like, dislike, or love sign done with the user's hand in front of the sensor.
We are providing a complete workflow from data acquisition to model training, then deployment on an STM32 NUCLEO-F401RE board. To create your end-to-end embedded application for the hand posture use-case, you simply need to follow these steps:
- Collect your custom dataset using the ST datalogging tool STSW-IMG035_EVK (Gesture EVK) by following this tutorial.
- Train a model on your custom dataset using the training tutorial.
- Alternatively, you can start directly from one of our pretrained models found in the models directory.
- Deploy the pretrained model with the deployment service.