BOS-HA - 基于深度学习的手语识别系统

BOS-HA是一个完整的手语识别系统，包括模型训练、实时识别、模型管理和Web客户端界面等功能。

功能特点

实时手语识别: 支持通过摄像头进行实时手语识别
多种模型支持: 支持PyTorch和OpenVINO模型
模型训练与转换: 内置训练模块，支持模型转换为OpenVINO格式
Web客户端界面: 提供直观的Web界面，支持实时显示和历史记录
WebSocket通信: 基于WebSocket的实时通信架构
可扩展的模型库: 支持多种模型的添加和切换

安装

pip install BOS-HA

快速开始

启动服务器

bosha-server

启动Web客户端

bosha-client

训练模型

bosha-train --config config.json

模型管理

列出可用模型

# 通过API获取
curl http://localhost:8000/models

切换模型

# 通过API切换
curl -X POST "http://localhost:8000/models/switch?model_name=hand_sign_model&model_type=pytorch"

技术架构

后端: FastAPI + WebSocket
前端: Flask + SocketIO + HTML/CSS/JavaScript
模型: PyTorch + OpenVINO
视频处理: OpenCV

目录结构

bosha/
├── server/         # 后端服务
│   ├── models/     # 模型文件
│   ├── services/   # 服务模块
│   └── main.py     # 主入口
├── client/         # 客户端
│   └── web_client/ # Web客户端
└── training/       # 训练模块

许可证

MIT License

贡献

欢迎提交Issue和Pull Request！

联系方式

项目主页: https://github.com/bos-ha/BOS-HA
联系邮箱: contact@bos-ha.com

Metadata

Release files for BOS-HA 1.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for BOS-HA 1.1.0
File	Size	Uploaded
bos_ha-1.1.0.tar.gz	72.7 kB	Dec 18, 2025	Details

Built distribution (wheel)

Table of built distributions (wheels) for BOS-HA 1.1.0
File	Interpreter	ABI	Platform	Reset
bos_ha-1.1.0-py3-none-any.whl 90.7 kB Dec 18, 2025	Python 3	none	any	Details

Total release size: 163.4 kB

Release files / bos_ha-1.1.0.tar.gz

Download URL	bos_ha-1.1.0.tar.gz
Size	72.7 kB
Tags	Source
SHA-256 checksum How to use checksums	`d5db83c5998a2621b0aa089eac70adde8b96131ee5f9145625219728e2c6ac33`
BLAKE2b-256 checksum How to use checksums	`94d9a27137a9fdba216899fe9a0475e38070f6201b3509c65660f5f854692190`
Upload date	Dec 18, 2025
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.11.9`

Release files / bos_ha-1.1.0-py3-none-any.whl

Download URL	bos_ha-1.1.0-py3-none-any.whl
Size	90.7 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`b3630df5af3531f532121c560d459937897a771642bd311938ae1085c2f1b0db`
BLAKE2b-256 checksum How to use checksums	`7df0fb2f1be2d2f6127cc964fdbd2ab9bd3ddf7013043cb568fc4132d34373a5`
Upload date	Dec 18, 2025
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.11.9`