Web Application

StreamPoseML includes a full-featured web application for real-time pose classification. The web application consists of:

• A React-based frontend for webcam capture and visualization

• A Flask-based API for classification

• Built-in MLflow integration for seamless deployment of trained models

This section covers the installation, configuration, and usage of the StreamPoseML web application.

MLflow Integration

A standout feature of the StreamPoseML web application is its direct integration with MLflow for model deployment. This provides several advantages:

• Standardized Model Serving: Deploy models tracked with MLflow without extra conversion steps

• Version Management: Easily switch between different model versions

• Metadata Tracking: Access model parameters, metrics, and artifacts

• Consistent API: Use the same interface for different model types

The web application includes a dedicated MLflow container that handles model loading and prediction requests, making it easy to deploy your trained models for real-time classification.

• Installation
  • Prerequisites
  • Quickstart Installation
  • Local Development Installation
  • Manual Docker Installation
• Usage
  • Accessing the Web Application
  • Web Interface Components
  • Model Configuration
  • Camera and Visualization
  • Bluetooth Device Integration
  • Performance and Optimization