WeAreDevelopers Logo WeAreDevelopers
Sign in
For You LIVE
WeAreDevelopers LIVE • Dec 7, 2021

Unboxing the DeepFace

Sefik Serengil

Standardizing eye alignment horizontally improves facial recognition accuracy by over 1%. Learn how the DeepFace framework abstracts complex computer vision pipelines into just a few lines of Python.

Pause
Mute Enter Fullscreen
#1 about 5 min

Introduction to the DeepFace Python library and setup

How to install open-source DeepFace and run basic verification tasks.

#2 about 2 min

Time complexity in face verification versus face recognition

Understanding the performance difference between one-to-one verification and one-to-many database searches.

#3 about 2 min

Extracting facial demography attributes like age and gender

Analyzing apparent demography and emotions to decrease search space and fight bias.

#4 about 3 min

Five core stages of a modern face recognition pipeline

Using powerful deep learning models to locate raw faces and landmarks.

#5 about 1 min

Aligning facial features to increase model recognition accuracy

Rotating input images to horizontally align eye coordinates prior to model ingestion.

#6 about 1 min

Normalizing facial images to remove noisy background information

Extracting just the facial visual data and discarding unhelpful background pixels.

#7 about 6 min

Generating vector embeddings with convolutional neural networks

Replacing traditional classification output layers with embedding arrays to support unknown future identities.

#8 about 4 min

Verifying face pairs by calculating multidimensional vector distances

Applying distance thresholding formulas to output embeddings to determine accurate matching identities.

#9 about 3 min

Preparing highly scalable facial recognition for big data

Reducing large-scale database search spaces through fast filtering by known demographic attributes.

#10 about 3 min

Speeding up similarity lookups with approximate nearest neighbor algorithms

Using heuristic indexing algorithms instead of linear neighbor comparisons to search billions of images.

#11 about 3 min

Exploring database architectures for storing facial vector embeddings

Storing resulting vectors efficiently in dedicated key-value stores rather than standard relational databases.

#12 about 3 min

Key advantages and flexibility of the DeepFace framework

Why lightweight design without complex C plus plus dependencies enables rapid deep learning integration.

#13 about 13 min

Answering audience questions on deepfakes and identical twins

Discussions regarding emotion accuracy, deepfakes, compatible sensor inputs, and recognizing identical twins.

Matching moments

5:25 min

Technical pipeline for real-time video deepfakes

Martin Förtsch Martin Förtsch +1 · World Congress 2024

3:09 min

Defining true deepfakes generated by trained neural networks

Thomas Endres Thomas Endres +2 · World Congress 2021

2:30 min

Implementing the deepfake inference workflow for face generation

Thomas Endres Thomas Endres +2 · World Congress 2021

3:36 min

Differentiating traditional video fakes from neural network models

Thomas Endres Thomas Endres +2 · World Congress 2021

3:32 min

Implementing automated object and face recognition for images

Arto Liukkonen · LIVE

3:59 min

Preparing video datasets and extracting faces with FaceNet

Thomas Endres Thomas Endres +2 · World Congress 2021

More from WeAreDevelopers LIVE

WeAreDevelopers LIVE - Project Chopin: Make Your Team and Agents Play Along thumbnail
1:21:17
6 days ago

WeAreDevelopers LIVE - Project Chopin: Make Your Team and Agents Play Along

Chris Heilmann, Daniel Cranney, Krzysztof Cieślak

Fake or News - AI Toothbrushes, Carrier Pigeon Tech & Life on Mars - Krzysztof Cieślak thumbnail
4:23
8 days ago

Fake or News - AI Toothbrushes, Carrier Pigeon Tech & Life on Mars - Krzysztof Cieślak

Chris Heilmann, Daniel Cranney

WeAreDevelopers LIVE - Open Source Agent Toolkits thumbnail
1:00:38
13 days ago

WeAreDevelopers LIVE - Open Source Agent Toolkits

Chris Heilmann, Daniel Cranney, Eddie Jaoude

Fake or News - AI Music Detectors, Translating Ancient Languages & Tiny Heads in Teslas - Eddie Jaoude thumbnail
6:16
15 days ago

Fake or News - AI Music Detectors, Translating Ancient Languages & Tiny Heads in Teslas - Eddie Jaoude

Chris Heilmann, Daniel Cranney

Fake or News: AI Agents Hire Each Other, Spider-Man Ads in Cars & RAM Thieves - Marc Plogas thumbnail
5:46
20 days ago

Fake or News: AI Agents Hire Each Other, Spider-Man Ads in Cars & RAM Thieves - Marc Plogas

Chris Heilmann, Daniel Cranney, Marc Plogas

WeAreDevelopers - Of Agents and Robots thumbnail
1:04:42
20 days ago

WeAreDevelopers - Of Agents and Robots

Chris Heilmann, Daniel Cranney, Marc Plogas

Upcoming sessions on this topic

Open session

World Congress 2026 North America

September 25, 2026 · 09:40–10:10

Stage 1

What's new, what's next: the latest models and developer tools from Google DeepMind

Paige Bailey

Member of the Technical Staff at Google DeepMind

Paige Bailey
Open session

World Congress 2026 North America

September 24, 2026 · 14:10–14:40

Stage 1

Anatomy of an AI Request: Where Latency and Cost Are Really Born

Dan Fu

VP of Kernels at Together AI

Dan Fu
Open session

World Congress 2026 North America

September 25, 2026 · 12:55–13:25

Stage 1

From Simulation to Reality: Overcoming the Data Scarcity Crisis in Physical AI

Mitesh Patel

NVIDIA Corporation, Developer Advocate -- Manager

Mitesh Patel
Open session

World Congress 2026 North America

September 23, 2026 · 10:00–17:00

Stage 11

Building Stuff with GenAI - The Open Minded Workshop beyond OpenAI

Andreas Erben

CTO for Applied AI and Metaverse at daenet

Andreas Erben
Open session

World Congress 2026 North America

September 24, 2026 · 14:10–14:40

Stage 7

Designing High-Performance AI APIs: Lessons from Serving Millions of Real-Time Requests

Wayne Liu

Chief Growth Officer and Americas President of Perfect Corp.

Wayne Liu
Open session

World Congress 2026 North America

September 25, 2026 · 11:40–12:10

Stage 9

You Can’t Re-Run Sunlight: Designing ML Data Architectures for Physical AI

An Phan

Senior Data Infrastructure Engineer @ Hippo Harvest

An Phan

Related videos

Deepfakes in Realtime - How Neural Networks Are Changing Our World thumbnail
46:42
World Congress 2021 • about 5 years ago

Deepfakes in Realtime - How Neural Networks Are Changing Our World

Thomas Endres, Martin Förtsch, Jonas Mayer

Getting Started with Machine Learning thumbnail
44:11
LIVE • almost 5 years ago

Getting Started with Machine Learning

Alexandra Waldherr

30 Golden Rules of Deep Learning Performance thumbnail
46:30
LIVE • about 6 years ago

30 Golden Rules of Deep Learning Performance

Anirudh Koul

Secure and Private AI - DeepMask thumbnail
4:57
World Congress 2025 • about 1 year ago

Secure and Private AI - DeepMask

Hissan Usmani

Focoos AI: Building the Future of Computer Vision thumbnail
4:18
World Congress 2025 • about 1 year ago

Focoos AI: Building the Future of Computer Vision

Antonio Tavera

Is it (F)ake?! Image Classification with TensorFlow.js thumbnail
25:43
World Congress 2025 • about 1 year ago

Is it (F)ake?! Image Classification with TensorFlow.js

Carly Richmond

Related articles

From learning to earning

Jobs that call for the skills explored in this talk.