One-Shot-Face-Recognition

One-Shot Face Recognition Using Siamese Neural Networks

In this work, we used convolutional neural networks (CNNs) to carry out the task of facial recognition.
More specifically, we have implemented a one-shot classification solution.
In order to achieve this goal, we implemented a Siamese Neural Network(SNN), according to the paper:
Siamese Neural Networks for One-shot Image recognition
Our main goal was to successfully execute a one-shot learning task for previously unseen objects.
Given two facial image of previously unseen persons, our SNN will try and determine whether they are the same person.

Sklearn API:
In order to use the dataset with sklearn, we have changed the function fetch_lfw_pairs:

1. Change the function that will support one channel picture
   • the change in sklearn\datasets\lfw.py --> new file of lfw was added to repo with comment added code
2. Redirector / Change to our dataset LFW-a instead of the existing one, and not download the matching dataset (LFW).

references:

1. Siamese Neural Networks for One-shot Image Recognition
2. Contrastive Loss