This project proposes a novel Face Recognition system employing One-Shot Learning, a technique tailored to discern facial similarities with minimal training data. Siamese neural networks are trained using both the Olivetti dataset and a custom dataset.
Dataset used for training: Enhanced Ollivetti dataset,150 epochs
The model is robust over changes in lighting , angles and noise
Model Architecture
Saved Model link to download:
https://drive.google.com/file/d/1Zrp-lf_mdUk98knedItTFaK8kLn7yXFG/view?usp=sharing
score above 0.5 correlates to the same person in the two images
