FER-Microservice

This repository contains a simple webservice offering an HTTP API to perform Facial Expression Recognition (FER).

POST a single image containing a human face and get a response with probabilities for seven basic emotions:

curl -X 'POST' \
  'http://localhost:8000/recognize/file' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'file=@fer/test/test_image.png'

{
  "anger": 0.051,
  "disgust": 0.049,
  "fear": 0.1,
  "happiness": 0.7,
  "neutral": 0.099,
  "sadness": 0.001,
  "surprise": 0.0
}

Getting Started

Docker

If you have Docker installed just build the image:

docker build -t fer-ms .

and run a container:

docker run -p 8000:8000 fer-ms

You may change the first 8000 to map the service to the desired port on your machine.

Python

You may find a complete exemplary machine/environment definition in this repository's Dockerfile.

If you just want to run the application, install all application requirements by running:

pip install -r requirements.txt

and then run the application via:

uvicorn fer.main:app --host 0.0.0.0 --port 8000

If you also want to run the tests, install all requirements by running:

pip install -r requirements-dev.txt

and then, to run the test:

python -m pytest fer/test

Example Usage

API via curl

You can find the API definition in main.py.

To GET a ping:

curl -X 'GET' \
  'http://localhost:8000/ping' \
  -H 'accept: application/json'

To POST an image file:

curl -X 'POST' \
  'http://localhost:8000/recognize/file' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'file=@fer/test/test_image.png'

To POST a base64 encoded image:

curl -X 'POST' \
  'http://localhost:8000/recognize/base64' \
  -H 'accept: application/json' \
  -H 'Content-Type: text/plain' \
  -d '@fer/test/test_image_base64.txt'

FER Live Demo

With the virtual environment activated, execute:

python demo.py

Tests in Python

Look around the fer/test folder for examples in Python code.