Skip to content

A deep learning-based gesture recognition system on ESP32-S3, featuring model training, quantization optimization, and resource-constrained deployment. Includes comprehensive workflow from dataset preparation to embedded implementation. 基于ESP32-S3的深度学习手势识别系统,涵盖模型训练、量化优化及资源受限部署的完整工作流程。包含从数据集准备到嵌入式实现的全过程。

Notifications You must be signed in to change notification settings

BlakeHansen130/esp32s3-gesture-dl

Repository files navigation

Gesture recognition system based on ESP32-S3 中文

A complete deep learning gesture recognition project, from model training to ESP32-S3 embedded deployment.

Project structure

  • training: Model training related files, including training code and various generated model files

  • quantization: Model quantization optimization, implementing multiple quantization strategies to adapt to embedded devices

  • deployment: ESP32-S3 deployment code, complete implementation based on ESP-IDF framework

  • dataset: Contains gesture image datasets for training and testing

  • environment: Environmental requirements (different steps require different environments).

Main features

  • Deep learning model for real-time gesture recognition
  • Support multiple quantization strategies to adapt to resource-constrained devices
  • Complete ESP32-S3 implementation based on ESP-IDF framework
  • Provide pre-trained models in multiple formats

Community Resources

  • Complete ESP-DL Workflow Guide
  • Comprehensive tutorial covering setup, model conversion, quantization and deployment
  • Environment setup
  • Model preparation
  • Quantization process
  • ESP32 deployment
  • Performance optimization

About

A deep learning-based gesture recognition system on ESP32-S3, featuring model training, quantization optimization, and resource-constrained deployment. Includes comprehensive workflow from dataset preparation to embedded implementation. 基于ESP32-S3的深度学习手势识别系统,涵盖模型训练、量化优化及资源受限部署的完整工作流程。包含从数据集准备到嵌入式实现的全过程。

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published