A Python package for face recognition.

Project Description

Gate6 Face Recognition Package

G6_face_recognition is a module for face recognition. Using the image processing libraries and high-level mathematical functions, we will provide fast and secure face recognition solution.

Installation

Installations Required:(Before installing the package module)

   python 
   numpy
   opencv-python
   matplotlib
   opencv-contrib-python
   requests
   cmake
   dlib
   scikit-image
   scipy
   imutils==0.5.2

- Install Python

Windows, Mac, Linux

- Install package module using pip:

  $ pip install G6-face-recognition

Project Structure

• In your project folder, create an encodingModel directory.
• In the encodingModel directory create a file named faceEncodings.pickle(encodingModel/faceEncodings.pickle).
• Create a directory named Input_database.
• In the Input_database directory put an individual's face images(in the directories made on their individual names).

    Project/
    ├── encodingModel/
    │   ├── faceEncodings.pickle/                               # train model
    |   ├── Face_model/                                         # Face model
    |       ├── face_recognition_model.dat/
    |       ├── shape_predictor.dat/
    ├── Input_database/ 
    │   ├── person1 name/                                       # person1 directory
    |   │   ├── face images of person1 /                    # images of person face
    │   ├── person2 name/                                       # person2 directory
    |   │   ├── face images of person2 /                    # images of person face
    │   ├── person3 name/                                       # person3 directory
    |   │   ├── face images of person3 /                    # images of person face                  

How to use

Once all the settings of project are configured, you are ready to run the project. Import G6_face_recognition module in your project to start.

   import G6_face_recognition

Once the import is completed, users need to copy face models from the link below and paste it in the directory Face_model.

   https://github.com/gate6/iris-recognition-sample-code/tree/face-recognition/encodingModel/Face_model

After that, users need to train existing images (which are saved in the Input_database Folder).

   Input_database/

Once it’s done, create and train encoding module using Input_database Folder images, as per the instructions given below:

   G6_face_recognition.face_model_train(train_database_path,train_encoding_model_path)
   train_database_path        ===>  Input_database/
   train_encoding_model_path  ===>  encodingModel/faceEncodings.pickle

Once the model is trained, it’s ready to test with real-time images. Follow the process that is mentioned to test real time face image:

   face_name = G6_face_recognition.face_model_test(test_encoding_model_path,real_time_image_path) 
   test_encoding_model_path   ===>  encodingModel/faceEncodings.pickle
   real_time_image_path       ===>  real-time_image_path
   face_name                  ===>  In response you’ll get the registered person’s name. If an image matches with the person’s image in the trained image model it will return as matched & if the image doesn’t match then the name returns as unmatched.

Requirements :

• Need clearer images from the input device.
• Images should be captured in light.
• Person should not wear big glasses or anything that affects their image match.
• Minimum 5 clear images are required to train the model.

Support

If you face any difficulty in configuration or while using our Gate6 Face Recognition Package (as per the instructions documented above), please feel free to contact our development team.

License

MIT