Its impossible to have thousands or millions of images of a person to train a Face Recognition model. Its better idea to use One Shot Learning to solve this problem. I used it and got reliable result.
Here, we will use Siamese Network to solve this problem.
Files to be used:
fr_utils.py as a functions to feed images to the network and getting the encoding of images.
inception_blocks_v2.py used to Compile FaceNet Network.
I got help from the following:
-
From Video Lectures from Deeplearning.ai by Andrew Ng.
-
FaceNet: A Unified Embedding for Face Recognition and Clustering : https://arxiv.org/abs/1503.03832
Thank You !!!