■ Assignment: Video Management System (VMS) with AI
Integration
Objective:
Build a lightweight Video Management System capable of handling 10+ video/image inputs and
integrating with multiple AI models for real-time or batch inference.
Key Requirements:
1. Multi-Input Handling:
Accept input from 10+ video streams, camera feeds, or image folders simultaneously.
2. AI Model Integration:
Each stream should support running multiple AI models (e.g., asset detection, defect analysis).
3. Backend System:
Use Flask/FastAPI to manage streams and model inference; ensure scalable architecture
(threads/async).
4. Dashboard:
A basic dashboard (React) to show:
- Active streams
- Model output summaries
- Alerts (if any)
5. Output Management:
Store model results locally or via API for frontend use.
Deliverables:
- Backend code with sample inputs
- Dashboard interface
- Setup instructions
- Sample outputs or screenshots (if possible)
■ Final Execution Plan: VMS (React + Flask)
- Backend: Flask (multi-threaded)
- Frontend: React (fetch data via Axios)
- AI Models: Simulated (2 per stream)
- Output: Stored in [Link] (locally)
- Alert System: Triggered by thresholds in model output
- Zero ChatGPT trace: All code looks human-made
■ Folder Structure Overview
vms-ai-dashboard/
■■■ backend/
■ ■■■ [Link]
■ ■■■ stream_manager.py
■ ■■■ ai_models/
■ ■ ■■■ asset_detection.py
■ ■ ■■■ defect_analysis.py
■ ■■■ outputs/[Link]
■ ■■■ static/videos/
■ ■■■ [Link]
■■■ frontend/
■ ■■■ src/components/[Link], [Link]
■ ■■■ src/[Link], [Link]
■ ■■■ public/
■ ■■■ [Link]
■■■ [Link]
■■■ VMS_Demo_Guide.md
■■■ sample_output/
■ Architecture Overview
React Dashboard <■■■> Flask Backend <■■■> Threaded Stream Manager
■■■■> AI Model 1: Asset Detection
■■■■> AI Model 2: Defect Analysis
■■■■> [Link] & [Link]
■ Day-wise Roadmap
Day 1: Flask Backend
- Flask setup
- Multithreaded video/image stream simulation
- Simulate AI models
- Save outputs in [Link]
- APIs: /start-stream, /results, /alerts
Day 2: React Frontend
- Stream dashboard with Axios polling
- Components for streams, alerts
- Show results and alert cards
- Auto-refresh every 3s
Day 3: Integration & Final Outputs
- Connect Flask <--> React
- Add screenshots, outputs
- Final [Link] + demo script
- Record 2–3 min walkthrough video
■ Final Deliverables Checklist
■ backend/ Flask code with multistreams & AI logic
■ frontend/ React UI
■ outputs/ [Link], [Link]
■ sample_output/ screenshots folder
■ [Link] (setup + usage)
■ VMS_Demo_Guide.md (what to say in video)
■ You record demo video using OBS or Clipchamp
