Person recognition using local binary patterns

software that detects specific persons using your webcam

config.json

• you can change the name of the person that you will be classifing
• you can change the number of samples ( pictures that are taken using 'collector.py' )

collector.py

• scans for faces using haarcascades and saves only the part where the face is

you need to create the folder 'faces' and inside it another folder named like in the config.json

train_and_detect.py

If you get a 'NoneType' error from the numpy.asarray function, make sure to delete the .DS_Store as the program sees it as None

• using the samples provided by collector.py i am using a local binary pattern histogram to detect a specific face
• after the classification was made, inside the window will be printed a confidence score
• it still doesn't provide the most accurate prediction but in the future i might use a cnn as the classifier