Deep Learning Specialization · Convolutional Neural Networks

Face Recognition with Triplet Loss

A FaceNet-style metric-learning project for face verification and recognition using 128-dimensional embeddings and triplet loss.

Triplet lossFace embeddingsOne-shot learningVerification vs recognitionL2 distance thresholding

GitHub

What this project is about

Face recognition is treated as metric learning rather than ordinary classification. The model maps a face to a 128-dimensional embedding, then compares distances for verification or recognition.

Implementation focus

This project clarified one-shot learning: instead of retraining a classifier for every new person, the system stores embeddings and compares them with a distance threshold.

Key results and artifacts

Embedding size: 128
Triplet margin: 0.2
Verification threshold: 0.7
Example match: younes with distance 0.5993

What I learned

This milestone showed how classification can be replaced by representation learning when the system must recognize people not seen during classifier training.

Milestone deep dive

What this project proves in my learning path.

A FaceNet-style metric-learning project for face verification and recognition using 128-dimensional embeddings and triplet loss. The milestone is included because it shows a complete learning unit: a problem, a data representation, an implementation path, measurable outputs, and a lesson that later connects to larger AI systems.

Input layercourse dataset, features, labels, or raw input artifacts

Model layerimplemented algorithm or neural network pipeline

Evaluation layertests, metrics, plots, or qualitative output

Artifact layermetrics, plots, predictions, and visual evidence

Problem framing

What I was trying to solve

This page is not just a lab description. It explains how I transformed the assignment idea into an engineering story: what the model receives, what it optimizes, what output it should produce, and how the result can be inspected.

Implementation

What I coded or assembled

The implementation focus was the mechanism behind the result: functions, model blocks, training logic, preprocessing, evaluation, and the small technical details that make the notebook work rather than only look correct.

Concepts

Core concepts

Triplet lossFace embeddingsOne-shot learningVerification vs recognitionL2 distance thresholding

These tags show the technical ideas this milestone added to my toolkit.

Learning result

What changed after it

This milestone showed how classification can be replaced by representation learning when the system must recognize people not seen during classifier training.

Important outputs and evidence

Embedding size: 128
Triplet margin: 0.2
Verification threshold: 0.7
Example match: younes with distance 0.5993

How I would explain it in an interview

I would describe this milestone by starting from the data representation, then explain the model or algorithm, then show what was measured. The important part is not only the final number; it is the ability to reason about why the method works, what assumptions it makes, and what limitations should be improved next.