oskar-k313/pyjudger

Image sharpness analyse using openCV

PyJudger

Falcon based REST API that main function is to analyze quality of face on given image.

Run Locally

PyJudger is docker based service, therefore all you got to do is build docker image and run docker-compose

Build docker:  
$ docker-compose build

Start server :
$ docker-compose up

by default it is running on 0.0.0.0:8313

Black

Project is formatted with Black

docker-compose run core sh -c "black core/*"

Available Endpoints

Sample body can be found in code/tests/filter_input.json

To generalize usage body requires only one attribute

{
  url: "www.someimagehosting.com/faceimage.jpg"
}

Sample images I have no rights to:

Blurry: https://i.ibb.co/b6GhhKL/blurry.jpg

Sharp: https://i.ibb.co/QY7zRjz/sharp.jpg

[http POST] /filter/all

    {  
      "var_laplacian": 52.40,
      "is_frontal": "FRONTAL",
    }

[http POST] /filter/isFrontal

    {
      "is_frontal": "FRONTAL"
    }

[http POST] /filter/laplacian

    {
      "var_laplacian": 800.01
    }

[http GET] /health

    {
     "message": "OK"
    }    

##TODO

[ ] Measure distance between eyes for better sharpness scaling

[ ] Allow thresholds to be passed in body

[ ] Divide image into smaller-blocks for better sharpness analysis

[ ] Logger

[ ] More tests!