main funding provided by infineon & de anza college
end-to-end facial recognition stack and control surface in one repo. train models, serve an api, and ship a next.js dashboard without juggling multiple projects.
- saba @sabaflz
- kevin
- charles
- jay
- ansel
- inky
- two-track monorepo –
backend/handles model training + inference;ix/ships the react 19 dashboard scaffold. - hardware ready – orange pi client scripts with infinite running capabilities (bash script + systemd service) for continuous monitoring.
- api first – flask rest service plus cli utilities for single-shot, health, and continuous recognition.
- frontend friendly – modern next.js (app router) starter ready to integrate live recognition results.
cd backend
python3 -m venv venv
source venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt- drop labeled images into
training/<person_name>/. - train models (writes
encodings.pkl+trained_folders.csv):python train.py
- boot the api (defaults to
http://localhost:5000):python api/app.py
cd sbc
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txtsingle-shot and health checks:
python orangepi_client.py # single capture
python orangepi_client.py health # ping the api
python orangepi_client.py continuous 5 # continuous monitoringinfinite running (production-ready):
-
bash script with auto-restart (recommended for testing):
./run_continuous.sh 5 # captures every 5 seconds, auto-restarts on crash -
systemd service (recommended for production):
# edit paths in orangepi-client.service first sudo cp orangepi-client.service /etc/systemd/system/ sudo systemctl daemon-reload sudo systemctl enable orangepi-client.service sudo systemctl start orangepi-client.service sudo systemctl status orangepi-client.service # check status sudo journalctl -u orangepi-client.service -f # view logs
extras
setup_orangepi.sh– one-liner bootstrap for orange pi hardware (installs dependencies, creates venv).run_continuous.sh– bash script with auto-restart for continuous monitoring.orangepi-client.service– systemd service file for production deployment.
cd ix-face-enroll
pnpm install # swap for npm/yarn if you prefer
pnpm dev- visit
http://localhost:3000to see the starter view inapp/page.tsx. - wire it up to the flask api (upload captures, poll
/health, display recognition events). - production helpers:
pnpm build pnpm start # serve the built app pnpm lint
| path | description |
|---|---|
backend/api/ |
flask rest api exposed for recognition and health checks. |
backend/training/ |
expected directory for per-person training images. |
sbc/ |
orange pi client scripts with infinite running capabilities. |
sbc/orangepi_client.py |
python client for capturing and sending images to api. |
sbc/run_continuous.sh |
bash script with auto-restart for continuous monitoring. |
sbc/orangepi-client.service |
systemd service file for production deployment. |
sbc/setup_orangepi.sh |
one-liner bootstrap script for orange pi hardware setup. |
model-train/ |
model training utilities and scripts. |
ix/ |
next.js 16 (react 19) app scaffold with tailwind-ready setup. |
- connect the dashboard to the api for upload/review flows.
- add automated smoke tests (api + client) before deploying to constrained hardware.
- package the backend with the provided dockerfile or integrate with ci/cd.



