Suriyakumarvijayanayagam
c6a2c392d9
Asked directly: "our cameras have an rtsp url, we can use that to connect them to this software right". Yes - and that has always been the mechanism, which is the point. CameraConfig.source() builds exactly that URL from the parts, and CameraConfig.url has always accepted a whole one and taken priority over them. No form ever offered it. So an operator holding the address their camera's own app shows had to split it into five fields by eye. That is where a password containing @ or / goes wrong, and this repository has already been bitten once by unencoded @ in RTSP credentials. parseRtspUrl lives in shared/cameraMakes.js and is imported by BOTH forms, for the same reason the make picker is: two copies would be worse than not offering it, because an operator trusts a filled-in field. A test asserts both import it. Decisions worth keeping: - Split into fields, not stored whole. Everything else on the form - Test, the make picker, editing later, and the rule that a password is never returned to the browser - works on the parts. A URL kept intact would carry the password back out to every screen that reads a camera. - WHATWG splits user info at the LAST @, which is what makes an unencoded @ inside a password parse the way a person means it. An operator doing it by eye would put "p" in the password box and "ssw0rd@192.168.1.121" in the address box. - Percent-encoded credentials are DECODED, because source() encodes again when it rebuilds the URL. Keeping them encoded would double-encode and the camera would refuse a password that is correct. - The scheme is optional, structure is not. Without requiring a slash, "nonsense" parses as a perfectly good hostname and silently fills the Address field with it - a wrong answer that looks like it worked. A bare address is refused too: the Address field already takes one. - A query string stays with the path. Some cameras carry the channel there, and dropping it opens the wrong channel - which looks like a camera pointed somewhere unexpected. - A URL carrying no credentials does not wipe a password already typed. Tested through node from pytest, the same pattern test_dashboard.py uses, and skipped when node is absent so the suite stays dependency-light. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KGcjxF1cNLcuwc3DAPcnfj
Behavision
Production face recognition over RTSP. Watches camera streams, detects and tracks faces, recognizes known people, auto-enrolls new visitors, records visit history, and serves a live dashboard + JSON API.
Clean-room rewrite of the previous Camera/ and pattern_reg/ projects:
same core ideas, correct engineering.
Quick start (Windows)
cd D:\NEARLE\Behavision
python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt
python -m behavision setup-models # downloads YuNet, copies ArcFace etc. from the old project
python -m behavision run # dashboard at http://localhost:8010
Camera credentials live in .env (gitignored) — never in code or YAML.
So do the dashboard credentials: set BEHAVISION_API_USER and
BEHAVISION_API_PASSWORD, or let the server generate one into
data/api_credentials.txt on first boot. A routable api.host is never
served without HTTP Basic auth; 127.0.0.1 is left open.
To test without a camera, set webcam: 0 on a camera in
config/default.yaml.
Enroll a person by name from photos:
python -m behavision enroll --name "Alice" --images C:\photos\alice\
Architecture
behavision/
├── config.py typed config: YAML + ${ENV} expansion, validated (pydantic)
├── capture.py RTSP/webcam reader thread: latest-frame slot, TCP transport,
│ exponential-backoff reconnect, percent-encoded credentials
├── detection.py YuNet face detector (OpenCV) → boxes + 5 landmarks, clipped
├── recognition.py ArcFace ONNX encoder (correct (x-127.5)/127.5 RGB preprocessing,
│ unit-norm output) + clamped face-quality scoring
├── tracking.py IoU tracker: identity decided once per TRACK, not per frame
├── gallery/
│ ├── index.py FAISS IndexFlatIP (exact cosine) with identical numpy fallback
│ ├── store.py SQLite (WAL): identities, embeddings, sightings — source of truth
│ └── service.py three-zone matching: match / ambiguous(do nothing) / enroll
├── attributes.py optional age, gender, emotion on the aligned chip
├── events.py async event bus → log / webhook / rate-limited email sinks
├── engine.py one worker thread per camera, shared models + gallery
├── api.py FastAPI: dashboard, MJPEG stream, identities, events, stats
└── __main__.py CLI: run | enroll | setup-models
Pipeline
RTSP ──► capture ──► detect (YuNet) ──► track (IoU)
│ once per track, quality-gated
▼
align (Umeyama 5-pt) ──► ArcFace ──► cosine search
│
┌───────────────────────────┼──────────────────────────┐
sim ≥ 0.42 0.32 ≤ sim < 0.42 sim < 0.32
known person ambiguous → retry new visitor
sighting + event on a better frame auto-enroll + event
Design decisions (and the failure they prevent)
| Decision | Prevents |
|---|---|
Per-camera FaceDetector, shared thread-safe encoder |
cv2 input-size race between camera workers |
| HTTP Basic on every route, escaped dashboard output | open biometric API on the LAN; stored XSS via identity labels |
| Percent-encoded credentials, URL built from parts | @ in password silently breaking the stream (old bug) |
| Track-level identity, sighting cooldown | one user registered per frame (old bug) |
Exact IndexFlatIP on unit vectors, -1 guarded |
inverted L2 threshold + wrong-person metadata[-1] (old bugs) |
| SQLite as source of truth, index rebuilt at boot | index/metadata drift, untrained-IVF crash (old bugs) |
| Three-zone thresholds with ambiguous no-op | duplicate identities and wrong merges |
| One color conversion, ArcFace-native normalization | off-distribution embeddings making thresholds meaningless (old bug) |
| All quality terms clamped to [0,1] | unreachable registration threshold (old bug) |
| Readiness-guarded API, sinks off the hot path | startup crashes, notification stalls |
API
| Method | Path | Purpose |
|---|---|---|
| GET | / |
live dashboard |
| GET | /api/health, /api/stats |
liveness / metrics |
| GET | /api/cameras/{id}/stream.mjpeg |
annotated live stream |
| GET | /api/cameras/{id}/frame.jpg |
latest annotated frame |
| GET | /api/identities, /api/sightings, /api/events |
data |
| PATCH | /api/identities/{id} |
rename a visitor ({"label": "Alice"}) |
| DELETE | /api/identities/{id} |
forget a person (embeddings removed) |
Tests
pip install pytest
pytest tests -q
Configuration
Everything lives in config/default.yaml; ${VAR} placeholders resolve
from the environment (.env is loaded first). Thresholds:
recognition.match_threshold(default 0.42): raise for fewer false matches, lower for fewer duplicates.recognition.min_enroll_quality(0.65): how good a face must look (sharpness, size, lighting, frontality) before a new identity is minted.tracking.min_hits_for_id(4): frames a face must persist before we spend an embedding on it — filters passers-by and phantom detections.cameras[].max_width(1280): frames are downscaled at ingest — full 3MP streams waste memory and detector time.
Recognition models
The encoder picks the first usable model in models/:
arcface_int8.onnx → w600k_mbf.onnx (MobileFaceNet, 13 MB, downloaded
automatically) → arcface.onnx (r100, 260 MB, copied from the old project;
needs ~1.5 GB free RAM to load). Pin one with recognition.model_file.
Every stored embedding is tagged with the model that produced it, and only
embeddings from the active model are searched — different encoders'
vectors are numerically incompatible and never mix.