Tinkerhub BFH-Project

This is an image classifier for popular malayalam actors lalettan and mammookka.
Uses ageitgey|face_recognition to detect faces and to build face encodings. Classifier used is SVM classsifier.
Pretrained model was trained using this dataset. Works on images that has both the actors in it.
Live link to test our Web App

Team members

1. Mohammed Aslam
2. Sagar Krishna
3. Muhammed Sabah

Team Id

BFH/recfJguBvrn88uh8W/2021

How it works

Step 1: face_recognition library will extract faces from image and face encoding data.
Step 2. The training data would be all the face encodings from all the known images and the labels are their names.
Step 3: svm classifier is trained with known face encodings and labels
Step 4. During testing an unknown face encoding is provided to the model Step 5. Output is obtained

Libraries used

Face recognition
sklearn
flask - for deployment dlib - for face_recognition

How to configure

Using pip

pip install -r requirements.txt

or

docker build .

How to run

For flask run :-

python app.py 

To use this to train your own model for other persons images

Make a directory containing images in the following order :

--directory_name
    |-----Images of person 1
    |-----Images of person 2
     ....
    |-----Images of person N

and run

python train.py directory_name

To test the model : First make a directory containing test images.

python test.py <modelname> <test_image_directory>

References

Face recognition
Develop a Face Recognition System