CV / ArcFace

ArcFace MobileFaceNet

reComputer R Series (CM5 + Hailo-10H)

ArcFace face recognition (2.04M params, 99.4% LFW) on reComputer R Series with Hailo-8 or Hailo-10H.

58 downloads

Size

3.91 MB

Memory

4GB+

Precision

Hailo HEF / HailoRT

Choose the device you're using, the set up guide and documentation will update accordingly.

Platform

Inference Engine

Getting Started

Deploy

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

Use the REST API to run inference. Copy the commands below.

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())

Model Details

reComputer R Series (CM5 + Hailo-10H)

ArcFace MobileFaceNet on reComputer R Series (CM5 + Hailo-10H)

ArcFace with a MobileFaceNet backbone extracts 512-dim face embeddings on Hailo-10H through HailoRT. The embedding is used for face verification and identification; face comparison itself is done by the client with cosine similarity.

This page targets reComputer R Series (CM5 + Hailo-10H) with a PCIe Hailo-10H accelerator.

Model Info

Property	Value
Architecture	ArcFace with MobileFaceNet backbone
Task	Face recognition (embedding extraction)
Input	112x112x3 RGB (letterboxed, normalization compiled into the HEF)
Output	512-dim face embedding
Parameters	2.04M
LFW accuracy	99.4% (Hailo Model Zoo reference)
HEF	Hailo Model Zoo v5.4.0, Hailo-10H

The accuracy value is Hailo's Model Zoo reference. It is not a benchmark measured on CM5.

Hardware and Host Setup

Item	Value
Board	reComputer R Series with Raspberry Pi CM5
Accelerator	Hailo-10H over PCIe, exposed as `/dev/hailo0`
Runtime	HailoRT 5.1.1
Python in container	3.13, aarch64

Install the Hailo-10H packages on the host and confirm that the driver and device are ready before starting the container:

bash

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

# After reboot
hailortcli --version
hailortcli fw-control identify
ls -l /dev/hailo0

Run With Demo Video

bash

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

Open http://<Board_IP>:8000 to view the web preview. Each frame is annotated with the embedding dimension, and the REST API returns the full 512-dim vector.

USB Camera Mode

bash

sudo docker run --rm \
  --name cm5-hailo-arcface \
  --privileged \
  --net=host \
  -e PYTHONUNBUFFERED=1 \
  --device /dev/hailo0:/dev/hailo0 \
  --device /dev/video0:/dev/video0 \
  -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 --camera_id 0

REST API

Prediction endpoint:

text

POST http://<Board_IP>:8000/api/models/arcface_mobilefacenet/predict

bash

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

json

{
  "success": true,
  "source": "uploaded image",
  "embedding": [0.023, -0.156, 0.089, "..."],
  "dimension": 512,
  "image": { "width": 1280, "height": 720 }
}

Endpoint	Method	Purpose
`/`	GET	Web preview UI
`/api/models/arcface_mobilefacenet/predict`	POST	Face embedding (JSON)
`/api/video_feed`	GET	MJPEG preview stream
`/api/video/upload`	POST	Upload a source video
`/api/video/analyze`	POST	Start offline video analysis
`/api/video/status`	GET	Read analysis progress
`/api/video/download/{filename}`	GET	Download the annotated result

Face verification is not exposed as a separate endpoint: compare two returned embeddings with cosine similarity on the client side (L2-normalize both vectors first).

HEF Files

HEF	Size	Notes
`arcface_mobilefacenet.hef`	2,666,496 bytes	Hailo-10H build, Model Zoo v5.4.0

Implementation Notes

The HEF contains its own normalization layer, so the application letterboxes frames to 112x112, converts BGR to RGB, and feeds raw uint8 pixels.
The embedding is read from the fc1 output when the vstream name is exposed; otherwise the first output tensor is used. The first inference logs every vstream name and shape so the mapping can be verified on hardware.
Empty or non-finite embeddings are rejected instead of being returned as a valid result.

Development Notes

Source module: src/hailo10h_arcface_mobilefacenet/
Dockerfile: docker/hailo10h/arcface_mobilefacenet.dockerfile
Container: ghcr.io/seeed-projects/recomputer-hailo10h-cv/arcface_mobilefacenet:latest
The module uses the HailoRT 5.1.1 InferModel API and a Hailo-10H-specific HEF.

Inputs and Outputs

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