The Ultimate Beginners Guide to Face Detection & Recognition

Name: The Ultimate Beginners Guide to Face Detection & Recognition
Rating: 4.5 (95 reviews)

Detect and recognize faces from images, videos and webcam using Python language with OpenCV and Dlib libraries!

Created byJones Granatyr, Gabriel Alves, AI Expert Academy

Last updated 4/2023

English

What you'll learn

Learn the differences between face detection and face recognition
Detect faces using Haarcascade, HOG (Histogram of Oriented Gradients), MMOD (Max-Margin Object Detection), and SSD (Single Shot Multibox Detector)
Detect and recognize faces in images, videos and from the webcam using OpenCV and Dlib libraries
Recognize faces using Eigenfaces, Fisherfaces, LBPH (Local Binary Patterns Histrograms), and advanced Deep Learing techniques
Evaluate face recognition algorithms in order to choose the best one according to your application

Course content

4 sections • 63 lectures • 8h 45m total length

Course content9:54
Detection vs recognition5:55
Course materials0:07

OpenCV and Dlib3:38
Images and pixels4:40
Haarcascade - intuition10:16
Loading the image9:56
Learn to load an image in Google Colab, convert it to grayscale, and run a higher cascade classifier to compare four face detection techniques, preparing for faster processing.
Face detection with haarcascade13:27
Learn to implement face detection with a haarcascade cascade classifier in OpenCV by loading the trained xml, detecting faces in grayscale, and drawing bounding boxes on the color image.
Resizing the image14:19
Haarcascade parameters 18:52
Haarcascade parameters 212:59
Tune mean neighbors and bounding box counts to balance detection quality and misses in Haarcascade face detection. Adjust mean size and scale factor to reduce false positives.
Eye detection11:21
Detection of smiles, clocks, body, and cars7:09
HOG (Histogram of Oriented Gradients) – intuition11:18
Face detection with HOG10:35
Upsampling parameter9:56
Explore upsampling in face detection with the lib library, showing how higher input resolution detects smaller faces and comparing the Haug algorithm with cascade classifiers.
Max-Margin Object Detection (MMOD) – intuition4:17
Face detection with MMOD10:57
Single Shot MultiBox Detector (SSD) – intuition6:37
Explore the single shot multi box detector, which localizes and classifies objects in one forward pass and regresses bounding box positions with confidences, achieving high mAP on Pascal VOC datasets.
Face detection with SSD 118:51
Face detection with SSD 24:41
HOMEWORK0:09
Homework solution6:42
Compare face detectors on a crowded image using haar cascade, hawk and svm, and sd. Tune parameters and image size to improve detections.
Face detection in videos 18:45
Detect faces in video frames using two techniques, higher cascades and SD, then resize the video, process frames, and create a new video with detections.
Face detection in videos 26:08
Implement video configurations and perform face detection using a higher cascade classifier, with grayscale processing and a video writer to output bounding boxes on each frame.
Face detection in videos 39:10
Face detection in videos 47:11
Face detection in videos 54:35
Installing Anaconda and PyCharm2:49
Webcam face detection6:04

Eigenfaces - intuition13:15
Yalefaces dataset9:57
Eigenfaces - implementation 115:51
Eigenfaces - implementation 212:51
Build a face dataset by extracting the region of interest, map images to subject IDs, train an OpenCV eigenfaces classifier, and save it as a yml file.
Eigenfaces - implementation 37:37
Eigenfaces - implementation 49:10
Eigenfaces - implementation 59:07
Fisherfaces - intuition3:41
Fisherfaces - implementation7:17
LBPH - intuition9:25
LBPH - implementation 13:41
LBPH - parameters4:36
LBPH - implementation 26:38
Learn how to implement lbp parameters for lbph recognition, adjust radius, neighbors, grid x and grid y, and threshold, and evaluate accuracy with results like 83%.
Deep learning with Dlib 111:15
Deep learning with Dlib 215:07
Learn to detect faces and 68 facial landmarks, extract face descriptors with the dlib face recognition model v1, and format a 128-number descriptor for each face in Colab.
Deep learning with Dlib 311:54
Deep learning with Dlib 49:28
Learn to compute face similarity with Dlib by measuring distances between face descriptors and identify the closest match by the lowest distance in a 138-face dataset.
Deep learning with Dlib 510:33
Deep learning with Dlib 67:09
Face recognition library 110:48
Face recognition library 210:33
Learn how to perform face recognition with multiple encodings using the face recognition library, including loading images, computing face encodings, detecting faces, comparing distances, and identifying best matches.
Face recognition library 314:30
Face recognition library 48:34
HOMEWORK0:14
Homework solution2:18
Apply the face detection and recognition algorithm to test images, using seven training images, read and display each image, and verify recognition results against labeled names to assess performance.
Face recognition library 59:25
Face alignment and face elements15:18
Video face recognition7:32
Apply video face recognition using a preloaded encoding file in Google Colab, detect faces, draw bounding boxes, and visualize results while noting when faces aren’t in the dataset.
Project: capturing faces via webcam11:29
Project: traditional algorithms6:47
Build a face recognition project using traditional OpenCV algorithms by creating a labeled image dataset, training Fisher faces and LBPH models, and testing recognition on webcam with various conditions.
Project: deep learning algorithms4:24

Requirements

Programming logic
Basic Python programming

Description

Facial detection is a subarea of Computer Vision that aims to detect people's faces in images or videos. Smartphones and digital cameras use these features to select people in a photo, usually placing a rectangle around the face. This type of application has gained considerable relevance in security systems, in which it is necessary to identify whether there are people in an environment for the alarm to be triggered. On the other hand, facial recognition aims to recognize people's faces and one example is security systems that can use these features to identify whether or not a person is present in an environment. It is important to highlight the differences between face detection and recognition techniques: while the first only indicates if a face is present, the second indicates whose face is detected.

In this step by step course using Python programming language, you are going to learn how to detect and recognize faces from images, videos and webcam from the most basic to the most advanced techniques! See below the topics that you be covered:

Detection of faces using Haarcascade, HOG (Histogram of Oriented Gradients), MMOD (Max-Margin Object Detection), and SSD (Single Shot Multibox Detector)
Detection of other objects, such as eyes, smiles, clocks, bodies, and cars
Recognition of faces using Eigenfaces, Fisherfaces, LBPH (Local Binary Patterns Histograms), and advanced Deep Learning techniques
How to compare the performance of the algorithms
Build your custom dataset capturing faces via webcam

All implementations will be done step by step using Google Colab online, so you do not need to worry about installing and configuring the tools on your own machine! More than 60 lectures and 8 hours of step by step videos!

Who this course is for:

People interested in face detection and face recognition using OpenCV and Dlib libraries
Undergraduate and graduate students who are taking courses related to Artificial Intelligence
Data Scientists who want to increase their project portfolio
Beginners in Computer Vision

The Ultimate Beginners Guide to Face Detection & Recognition

What you'll learn

Explore related topics

Course content

Introduction3 lectures • 16min

Face detection27 lectures • 3hr 45min

Face recognition31 lectures • 4hr 40min

Final remarks2 lectures • 4min

Requirements

Description

Who this course is for: