Combining the popular face_recognition library with a simple flask backend that can receive images and respond with the labeled ones, without saving them locally.

pip3 install -r requirements.txt

• To run the server with the sample model

python3 api.py

• To train your own KNN classifier with your dataset, place your images according to the structure below:

images/train/
        ├── <person1_name>/
        │   ├── <someimage1>.jpeg
        │   ├── <someimage2>.jpeg
        │   ├── ...
        ├── <person2_name>/
        │   ├── <someimage1>.jpeg
        │   └── <someimage2>.jpeg
        └── ...

python3 train.py

The trained model will be placed in the root folder, named as "trained_knn_model.clf"

• Send POST requests to the /upload endpoint with a file on them as form-data

MIT