- This is a simple Face Recognition project using Python OpenCV, to learn and understand the basics of a project. This project is inspired by https://github.com/leodlca/lbph-face-recognition.
- For Face Detection we use OpenCV's cascade: haarcascade_frontalface_alt.xml, which can be found at https://github.com/opencv/opencv/blob/master/data/haarcascades/haarcascade_frontalface_alt.xml.
- We use the Local Binary Pattern Histogram (LBPH) Algorithm for training and recognizing the different faces. LBPH algorithm creates More about the LBPH algorithm at https://towardsdatascience.com/face-recognition-how-lbph-works-90ec258c3d6b
We use the OpenCV library's pre-trained models to perform the task of Face Recognition using the LBPH algorithm. OpenCV offers several for the task of object Detection. We use the Frontal-Face Haar Cascade to detect a "face" in the frame. Once a face is detected it has a bounded box to find its location, and the face is extracted, leaving aside the other non-important details in the frame. The LBPH algotihm is then run on the extracted face.
- Python 3.6+
- OpenCV
- Numpy
- Pandas
- Run
python TakePhotos.py - Enter an ID and Name
- Press the
skey repeatedly to take photos, once a box appears around your face. It is recommended to take atleast 25 pictures. - Press the
qkey when you're finished taking pictures.
- Run
python Train.py - After Training is complete the program will generate the file "Classifiers/TrainedLBPH.yml"
- Run
python Recognize.py