Let’s Recognize Face Using Transfer Learning

Hey guy’s In this article, I going to show you how to do Face Recognition using pre-trained architecture like VGG16 by the process of Transfer Learning.

But Before we go further we need to have following library installed in your system TensorFlow, Keras, Numpy, OpenCV-Python, pillow and pandas.

In this article I am using Jupyter notebook But if you want you using spyder or pycharm but install above library as mentions.

Hear I show how to install library. Because I am using anaconda hence I use “ conda install numpy ” but if you are using pycharm then use “pip”

Now get the image of at least 4 different people but if you want you can get more. that collected image we called dataset.

For further use we divide the collected data into the 80/20. We use 80 % set in the training and rest for the testing. Now for the convenience I am dividing the data and putting in different folder. One folder for the testing and another one for training.

Now let come to very important part. In the face recognition we are using Pre-trained Model VGG16 which is already present in keras framework.

Guy’s Now I am including some pic for the better understanding.