Multi-Face-Comparison

Compare two images and it will return True or False. This is the main focus of this repo. I have attached two py file. One is compare_image.py this can be run in the terminal with two images name like

karthick_aravindan@ml-machine:~/Documents/Two-Face-Comparison$ python compare_image.py image_1.jpg image_2.jpg

Another File is app.py. This file is api which can be attached to website or other apps. In this we compare multiple faces with one target image.

Requirment

pip install face_recognition
pip install flask
pip install scipy

Run

> karthick_aravindan@ml-machine:~/Two-Face-Comparison$ python app.py

Flask api will be hosted 8000 port

Endpoints

• /api/v1/compare_faces
• /api/v1/detect_faces

Output for compare_image.py

Distance and Boolean value. Distance is higher means it is false and lower is True

command line output

> karthick_aravindan@ml-machine:~/Two-Face-Comparison$ python compare_image.py  ami.jpg AMI.jpg  
> 0.5439451765528829
> True

Output for compare face API

Output will be in json format. Addition I have calculate time take to predict

Example

[
    {
        "result": "True",
        "distance": 0.0,
        "time_taken": 0.29,
        "target": "rajini.jpeg",
        "face": "rajini.jpeg"
    },
    {
        "result": "False",
        "distance": 0.77,
        "time_taken": 0.339,
        "target": "rajini.jpeg",
        "face": "kamal.jpg"
    }
]    

I have used postman to check the api. The url mostly will be like http://localhost:8000/compare_faces in POST format.

form-data key name will be target and faces in Postman

Detect Faces API

It will returns list of face coordinates for the images.

Example

[
    {
        "coordinates": [
            {
                "ymin": 46,
                "xmin": 175,
                "ymax": 136,
                "xmax": 86
            }
        ],
        "time_taken": 0.046,
        "image_name": "kamal.jpg"
    },
    {
        "coordinates": [
            {
                "ymin": 32,
                "xmin": 101,
                "ymax": 84,
                "xmax": 49
            }
        ],
        "time_taken": 0.016,
        "image_name": "harry.jpg"
    }
]

command line output

> karthick_aravindan@ml-machine:~/Two-Face-Comparison$ python detect_face.py image_path/ami.jpg
> Coordinates:  [(53, 114, 115, 52)]

Links for other resources

https://face-recognition.readthedocs.io/en/latest/installation.html