go-faces-http

Face detection HTTP microservice based on dlib.

Installation

Portable statically-linked binary for Linux AMD64 is available on releases.

wget https://github.com/AndriyKalashnykov/go-faces-http/releases/download/latest/linux_amd64.tar.gz && tar xf linux_amd64.tar.gz && rm linux_amd64.tar.gz
./faces -h

It is also available as a docker image.

docker run --rm -p 8011:80 ghcr.io/andriykalashnykov/go-faces-http:v0.0.5

If you want to build the app from source, please follow the instructions on dependency setup.

Usage

./faces -h
Usage of ./faces:
  -listen string
        listen address (default "localhost:8011")

Start server.

./faces
2024/01/15 23:44:22 recognizer init 424.357089ms
2024/01/15 23:44:22 http://localhost:80/docs

Send request.

curl -X 'POST' \
  'http://localhost:8011/image' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'image=@faces.jpg;type=image/jpeg'

{
  "elapsedSec": 2.373028184,
  "found": 4,
  "faces": [
    {
      "Rectangle": {
        "Min": {
          "X": 584,
          "Y": 1228
        },
        "Max": {
          "X": 1029,
          "Y": 1673
        }
      },
      "Descriptor": [
        -0.122200325,
        0.10511437,
        0.05358115,
        0.011272355,
        -0.09460048,
        "............. cut here ..........."
      ]
    }
  ]
}

This repo contains models, that were created by Davis King <https://github.com/davisking/dlib-models>__ and are licensed in the public domain or under CC0 1.0 Universal. See LICENSE.

References

• Building a platform-specific executable with Docker Image and uploading to GiHub releases