FallDetection Platform

Multi-service architecture for a fall detection scenario:

• FallDetectionAPI/: .NET API + SignalR
• ai-service/: FastAPI-based AI inference service
• mock-camera-service/: MediaMTX + FFmpeg test camera streaming
• factory-dashboard/: React dashboard (HLS monitoring + event panel)
• bruno_falldetection/: API test collection

Quick flow

1. mock-camera-service publishes test video streams via RTSP/HLS.
2. FallDetectionAPI manages camera inventory and event flow.
3. ai-service is called by the API to produce predictions.
4. factory-dashboard shows video and events through API + SignalR + HLS.

Default ports

Service	Port	Compose path
WebAPI	5000	FallDetectionAPI/docker-compose.yml
SignalR	5001	FallDetectionAPI/docker-compose.yml
AI	8200	ai-service/docker-compose.yml
MediaMTX	8554 (RTSP), 8888 (HLS)	mock-camera-service/docker-compose.yml
Dashboard	3000	factory-dashboard/docker-compose.yml

Run services with Docker

There is no single central compose file; each service is started from its own folder.

First, create the shared network once:

docker network create fall_detection_network

Then use this typical startup order:

1. Mock camera

cd mock-camera-service && docker compose up -d --build

2. AI service (a token may be required for initial model download)

# PowerShell
$env:HF_TOKEN="hf_xxx"
cd ai-service && docker compose up -d --build

3. API + SignalR + Postgres

cd FallDetectionAPI && docker compose up -d --build

4. Dashboard

cd factory-dashboard && docker compose up -d --build

Access:

• Dashboard: http://localhost:3000
• API: http://localhost:5000
• SignalR: http://localhost:5001
• HLS: http://localhost:8888

Folder structure

FallDetection/
├── FallDetectionAPI/        # .NET API + SignalR + layered backend
├── ai-service/              # FastAPI-based AI service
├── mock-camera-service/     # RTSP/HLS test streaming infrastructure
├── factory-dashboard/       # React dashboard
└── bruno_falldetection/     # Bruno API test collection