/face-detection-tutorial

Repository to demonstrate some face detection features from Microsoft Azure Cognitive Services

Primary LanguagePythonMIT LicenseMIT

Face detection using Azure Cognitive Services (Face API)

This implementation is for learning purposes. The class face.py could be used to train Cognitive models using Face API. Could be used in scenarios like person identification, fraud detection, and much more. Use your creativity to expand the use to have fun to fix a lot of real world problems and feel free to clone this repo and adapt this code for your requirements :)

Activate environment

pipenv shell

Flake8 config

Use the bellow code to adjust the max line length limit flake8 --max-line-length=140

If you use VSCode you can add this config to settings.json adding the code:

"python.linting.flake8Args": [
        "--max-line-length=140",
    ]

Set environment variables

You can find API Key in Quick start section of your API on Azure Portal

export FACE_SUBSCRIPTION_KEY=<KEY_1>

You can find API endpoint in Quick start section of your API on Azure Portal

export FACE_ENDPOINT=<FACE_ENDPOINT>

How to train the model

Run run_train.py and change the variables PERSON_GROUP_ID and the parameters person_name and faces_image_path for the method __assign_person_image_to_person_group__

PERSON_GROUP_ID: is the name for the person group to be created (eg. People, Animals, etc)

person_name : is the name of the person (Eg. Barack Obama)

faces_image_path: is the path of the images to be used for training (Eg. obama_1.jpg, obama_2.jpg). You should download all the images associated with the person and put all of them inside the folder used in the path

How to identify a person associated with an image

Run run_identify.py and change PERSON_GROUP_ID and FACES_IMAGE_PATH_TO_BE_IDENTIFIED. It is important to use a PERSON_GROUP_ID that was already trained before (see How to train the model section).

PERSON_GROUP_ID: is the name for the person group already used for the training stage FACES_IMAGE_PATH_TO_BE_IDENTIFIED: is the path of the images to be identified. You can also use the is_url flag of the method identity_face to be able to identify images directly from the web.

Clean up resources

To clean up resources you can delete the Resource Group or the API using Azure Portal. If you don't want to lose the resources you can also run the file run_cleanup.py this method only delete the person_group_id.