Face recognition with Convolutional Neural Network Martin Vels - - PowerPoint PPT Presentation

▶
face recognition with convolutional neural network
SMART_READER_LITE
LIVE PREVIEW

Face recognition with Convolutional Neural Network Martin Vels - - PowerPoint PPT Presentation

Oct 17, 2023 29 likes •159 views

Face recognition with Convolutional Neural Network Martin Vels Face recognition with CNN Labeled Faces in the Wild (LFW) dataset with 13,233 images, 5749 persons (classes) Only using classes with 5 or more samples: 423 classes

slide-1
SLIDE 1

Face recognition with Convolutional Neural Network

Martin Vels

slide-2
SLIDE 2
  • Labeled Faces in the Wild (LFW) dataset

with 13,233 images, 5749 persons (classes)

  • Only using classes with 5 or more

samples: 423 classes

  • Using Convolutional Neural Network

(CNN) to recognize person on the image

Face recognition with CNN

slide-3
SLIDE 3
  • Face recognition and in general pattern

recognition are interesting topic

  • My research is related to analyzing video

data to find certain patterns

  • Video is a sequence of images
  • Get to know the topic of CNN and use the

knowledge in my research

Motivation

slide-4
SLIDE 4
  • CNN can achieve really good results on

image data

  • Sample CIFAR-10 dataset with 60k

images and 10 classes achieves <2% error rates

  • With LFW dataset, achieving 30% error

rate would be reasonable

Goal

slide-5
SLIDE 5
  • 423 classes, 5985 images,
  • median number of images per class: 8,

50% of classes with 5 images,

  • most images per class: 530
  • image size 250x250px
  • cropping 128x128 from center, resizing to

64x64px

  • some experiments with grayscale images

LFW Dataset

slide-6
SLIDE 6

LFW Dataset - resizing

slide-7
SLIDE 7
  • Similar to regular neural network
  • Basic building block is neuron
  • Neurons are organized into layers
  • Various types of layers
  • Idea is to gradually reduce high

dimensional input and classify the image

Convolutional Neural Network

slide-8
SLIDE 8

http://cs231n.github.io/neural-networks-1/

Neuron - the main building block

slide-9
SLIDE 9

Various types of layers to reduce dimensions

http://cs231n.github.io/convolutional-networks/

ConvNet architecture

slide-10
SLIDE 10
  • MATLAB toolbox
  • Implements Convolutional Neural

Networks for computer vision applications

  • CNN building blocks available as functions
  • Available freely: http://www.vlfeat.
  • rg/matconvnet/

MatConvNet

slide-11
SLIDE 11

Results

10 classes with at least 50 images per class 32x32 RGB 64x64 grayscale

slide-12
SLIDE 12

Results

423 classes with mostly less than 8 images per class 32x32 RGB 64x64 grayscale

slide-13
SLIDE 13
  • CNN is an interesting and promising tool
  • Works well with large dataset
  • Disappointing results with my dataset

Future ideas:

  • Use horizontal flipping and cropping from

corners to generate more data

  • Experiment with different configurations

and parameters

Conclusion