Spaces:
Sleeping
Sleeping
metadata
title: FocusFlow
emoji: π―
colorFrom: blue
colorTo: indigo
sdk: docker
pinned: false
FocusFlow: Real-Time Meeting Engagement Analytics
Premium Engagement Tracking Powered by AI and C++
Focused on Professional Productivity and Group Dynamic Monitoring.
π Project Overview
FocusFlow is a modern web application that analyzes video feeds (camera or screen share) to compute real-time Engagement Scores. It provides deep insights into focus levels, emotional state, and group dynamics during professional meetings.
Key Features:
- Computer Vision: Face detection, head pose estimation, and iris tracking via MediaPipe.
- C++ Performance: High-speed metrics calculation via custom C++ PyBind11 extension.
- Meeting Intel: Multi-participant detection and group engagement analytics.
- Modern UI: Asymmetric, professional dashboard design with real-time WebSocket feedback.
- Persistence: SQLite-backed session history and trend analysis.
ποΈ Architecture
graph TD
A[Frontend: Vanilla JS] -->|Frames| B[FastAPI WebSocket]
B -->|Base64| C[Vision Engine]
C -->|Points| D[C++ Module]
D -->|Score| C
C -->|Metrics| B
B -->|JSON| A
π οΈ Tech Stack
| Component | Technology |
|---|---|
| Backend | Python 3.10+, FastAPI |
| Frontend | HTML5, CSS3, Vanilla JavaScript |
| Vision | OpenCV, Google MediaPipe |
| Performance | C++17, PyBind11 |
| Database | SQLite3 |
π Clean Folder Structure
FocusFlow/
βββ cpp_modules/ # C++ Performance Module (Source)
βββ frontend/ # Modern Web Interface (HTML/CSS/JS)
βββ src/ # Core Application Logic
β βββ main.py # FastAPI Gateway & Websockets
β βββ vision_engine.py# AI Vision Orchestrator
β βββ database.py # SQLite Persistence Layer
β βββ utils/ # Utility Scripts
βββ data/ # Storage (Database files)
βββ architecture.md # Detailed Technical Docs
βββ file_manifest.md # π Breakdown of every file
βββ run.bat # One-Click Launcher (Windows)
βββ build_cpp.bat # C++ Compilation Tool
βββ requirements.txt # Dependencies