hand_posture

ST multi-zone Time-of-Flight sensors hand posture recognition STM32 model zoo

Directory components:

• datasets placeholder for the Hand Posture ToF datasets.
• deployment contains the necessary files for the deployment service.
• pretrained_models a collection of optimized pretrained models
• src contains tools to train, evaluate and benchmark your model on your STM32 target.

Quick & easy examples:

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 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 readme links below that provide typical examples of operation modes, and tutorials on specific services:

• training
• evaluation
• benchmarking
• deployment

Guidelines

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 posture such as a like, dislike or love sign done with user 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 ST datalogging tool STSW-IMG035_EVK (Gesture EVK), by following this tutorial.
• Train a model on your custom dataset using the training scripts provided here.
• Alternatively, you can start directly from one of our pretrained models found in the models directory.
• Deploy the pretrained model with the deployment service.