This repository is based on a project completed as part of the Deep Learning Specialization on Coursera by DeepLearning.AI.
The objective of this project is to construct a facial recognition system inspired by FaceNet.
Facial recognition tasks are typically categorized into two main types:
-
Face Verification - This involves confirming if the presented individual matches the claimed identity, a one-to-one (1:1) matching problem.
-
Face Recognition - In this scenario, the objective is to identify the person in the image, a one-to-many (1:K) matching problem.
FaceNet utilizes a neural network to encode a facial image into a 128-dimensional vector. By comparing two such vectors, it becomes possible to determine whether two images depict the same individual.
In this assignment, the following items were explored:
- Implement the triplet loss function
- Use a pretrained model to map face images into 128-dimensional encodings
- Use these encodings to perform face verification and face recognition