CV / ArcFace

ArcFace MobileFaceNet

ArcFace with MobileFaceNet backbone extracts 512-dim face embeddings for face verification and identification through HailoRT inference.

2 downloads

Grootte

2.5 MB

Geheugen

4GB+

Precisie

Hailo HEF / HailoRT

Kies het apparaat dat je gebruikt. De installatiehandleiding en documentatie worden dienovereenkomstig bijgewerkt.

Platform

Inferentie-engine

Aan de slag

Implementeren

sudo docker run --rm \
  --name cm5-hailo-arcface \
  --privileged \
  --net=host \
  -e PYTHONUNBUFFERED=1 \
  --device /dev/hailo0:/dev/hailo0 \
  -v /usr/lib/libhailort.so.5.1.1:/usr/lib/libhailort.so.5.1.1:ro \
  -v /usr/lib/libhailort.so:/usr/lib/libhailort.so:ro \
  ghcr.io/seeed-projects/recomputer-hailo10h-cv/arcface_mobilefacenet:latest \
  python web_detection.py --model_path model/arcface_mobilefacenet.hef --video_path video/test.mp4

REST API

Gebruik de REST API om inferentie uit te voeren. Kopieer de onderstaande commando's.

Curl

curl -X POST "http://<Board_IP>:8000/api/models/arcface_mobilefacenet/predict" \
  -F "file=@face.jpg"

Python

import requests
resp = requests.post(
    "http://<Board_IP>:8000/api/models/arcface_mobilefacenet/predict",
    files={"file": open("face.jpg", "rb")},
    timeout=30,
)
print(resp.json())

Modeldetails

ArcFace MobileFaceNet on reComputer CM5 + Hailo-10H

ArcFace + MobileFaceNet (2.04M params, 99.4% LFW) for face recognition, running on Hailo-10H.

Model Info

Property	Value
Architecture	ArcFace + MobileFaceNet
Task	Face Recognition
Input	112×112×3 RGB
Output	512-dim embedding
Parameters	2.04M
LFW Accuracy	99.4%
Source	Hailo Model Zoo

Hardware and Host Setup

Item	Value
Board	Raspberry Pi CM5
Accelerator	Hailo-10H, `/dev/hailo0`
OS	Raspberry Pi OS Bookworm, aarch64

bash

sudo apt update && sudo apt install hailo-h10-all -y
sudo reboot

Run

bash

sudo docker run --rm --privileged --net=host \
  -e PYTHONUNBUFFERED=1 \
  --device /dev/hailo0:/dev/hailo0 \
  -v /usr/lib/libhailort.so.5.1.1:/usr/lib/libhailort.so.5.1.1:ro \
  -v /usr/lib/libhailort.so:/usr/lib/libhailort.so:ro \
  ghcr.io/seeed-projects/recomputer-hailo10h-cv/arcface_mobilefacenet:latest \
  python web_detection.py --model_path model/arcface_mobilefacenet.hef --video_path video/test.mp4

HEF Files

HEF	Size	Notes
`arcface_mobilefacenet.hef`	2.5 MB	Face embedding model

REST API

Extract Embedding

bash

curl -X POST "http://<Board_IP>:8000/api/models/arcface_mobilefacenet/predict" \
  -F "file=@face.jpg"

The response contains a 512-dimensional embedding and its dimension. Face comparison is performed by the client; this service does not expose a separate compare endpoint.

Development

Source: src/hailo10h_arcface_mobilefacenet/
Dockerfile: docker/hailo10h/arcface_mobilefacenet.dockerfile

Invoer en uitvoer

Input: image (112x112 RGB). Output: 512-dim face embedding vector.