Scribd
Upload
259 views16 pages

Face Recognition Using LBPH

This document summarizes face recognition using Local Binary Pattern Histograms (LBPH). It discusses how LBPH can extract local features from faces and represent them as histograms for comparison. The key steps are face detection using an LBPH cascade classifier, image resizing, extracting LBPH features to create histograms, and comparing histograms to recognize faces. LBPH is computationally simple and fast, and robust to illumination changes and occlusion. While less accurate than other methods, it provides real-time face detection and recognition capabilities.

Uploaded by

Shailesh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
259 views16 pages

Face Recognition Using LBPH

This document summarizes face recognition using Local Binary Pattern Histograms (LBPH). It discusses how LBPH can extract local features from faces and represent them as histograms for comparison. The key steps are face detection using an LBPH cascade classifier, image resizing, extracting LBPH features to create histograms, and comparing histograms to recognize faces. LBPH is computationally simple and fast, and robust to illumination changes and occlusion. While less accurate than other methods, it provides real-time face detection and recognition capabilities.

Uploaded by

Shailesh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 16

Face Recognition Using LBPH

Presented By:
Shailesh Arya 16BCP002
Hrithik Mesariya 16BCP012
Kaushal Kapadiya 16BCP018
Vedant Bhavsar 16BCP066
Akshay Modi 16BCP078
Niharika Chavda 16BCP080D
Introduction
• The face of the human is a very complex structure
• Humans can identify faces with brain but it’s quite difficult for
computer system
• Human face changes due to aging, spectacles , beard, hair style etc.
• Involves feature extraction, face detection and face recognition.
• The recognition efficiency depends on training and testing images
selected.
• Commonly used for Security purpose
• Applications: Airport, criminal detection, face tracking, forensic etc.
Problem Statement
• Locate and Identify Human Faces in an image regardless of their-
• Pose
• Location
• Orientation
• Illumination
• Recognize the face
• By comparing with the dataset
Different Approaches
 Eigen faces  Fisher Faces  Linear Binary Pattern
• This method is focusing • Improved version of • Overcomes the
on the areas of Eigen face drawback of the Eigen
maximum change faces and Fisher faces.
• Lower Error rates
i.e. variance of face. • Local feature of an
This uses PCA to • Looks all training faces image are taken into
identify the important at once and finds PCA consideration
component. combined • Finds local structure of
• Drawbacks: it is an image
affected by light. • by comparing each
pixel with its
neighboring pixels.
Approach
• Pre-processing
• Convert image to Grayscale
• Face Detection
• Using LBP cascade classifier
• Image Resizing
• Face Recognition
• Create Histogram
• Comparison of Histogram
Face Detection Using LBP cascade Classifier
• LBP (Linear Binary Pattern) is a visual/texture descriptor and
thankfully, our faces are also composed of micro visual patterns
• Steps:
• Images are converted into gray scale image.
• LBP features are extracted to form a feature vector that classifies a face from
a non-face.
Working of LBP
• Each training image is divided into blocks
• 9 Pixels (as 3X3 matrix )

Source: https://bit.ly/2UxY869
• At the end, we have a new image which represents better the
characteristics of the original image
Working of LBP
• The LBP procedure was expanded to use a different number of radius
and neighbors, it is called Circular LBP.

Source: https://bit.ly/2UxY869

• It can be done using bilinear interpolation.


• If some data point is between the pixels, it uses the values from the nearest
pixels(2x2) to estimate the value of the new data point.
Working of LBP
• Each block values are converted into a Histogram
• One histogram per Image
Face Recognition Using LBPH
• On new input: Create a Histogram for that Image
• Compare Histogram with already available Histogram & return the
closest Match.
• Different Methods for comparing Histogram
• Calculate distance between 2 histogram
• example : Euclidean Distance, Chi-Square, Absolute Value etc.
• Output of Algorithm:
• ID of the image with closest Histogram
• Calculated Distance as confidence measurement
Conclusion
• LBPH is one of the easiest face recognition algorithms.
• It can represent local features in the images.
• Possible to get great results (mainly in a controlled environment).
• Robust against monotonic gray scale transformations.
• Faster than Eigen face and Fisher face.
• Better works in Real time face detection and recognition
• Used in texture classification
Conclusion
• Overcomes the drawback of the Eigen faces and Fisher faces.
• (i.e. it is not affected by illumination. )

Source: https://bit.ly/2I8PJ23
LBP vs HAAR Classifier
Algorithm Advantages Disadvantages

1. Computationally complex
and slow
2. Longer training time
1. High detection Accuracy
HAAR 3. Less accurate in black faces
2. Low False Positive Rate
4. Limitations in difficult
lightning conditions
5. Less robust to occlusion
1. Computationally simple and
fast
1. Less accurate
LBP 2. Shorter training time
2. High false positive rate
3. Robust to local illumination
4. Robust to occlusion
Future Work
• Facial Expression
• Real time application
References
• https://ieeexplore.ieee.org/document/1717463
• http://ripublication.com/acst17/acstv10n5_47.pdf
• https://www.irjet.net/archives/V4/i6/IRJET-V4I6800
• http://www.scholarpedia.org/article/Local_Binary_Patterns
• https://towardsdatascience.com/face-recognition-how-lbph-works-
90ec258c3d6b
• https://www.superdatascience.com/blogs/opencv-face-recognition
• https://www.superdatascience.com/blogs/opencv-face-detection/
• https://en.wikipedia.org/wiki/Local_binary_patterns
THANK YOU

You might also like