21
Solving simple problems using Neural Network Mark Nguyen

Solving Simple Problems With Neural Networks presented by Mark Nguyen

Embed Size (px)

Citation preview

Solving simple problems using Neural Network

Mark Nguyen

Neural Network

ContentsI. Face recognition using single layer neural network

1. Database

2. Source code

3. Result

II. Face recognition using deep neural network

III. Spoof fingerprint detection using convolution neural network

Face Recognition using single layer NN• ten different images of each of 40 distinct

subjects• the images were taken at different times,

varying the lighting, facial expressions and facial details

• the size of each image is 92x112 pixels, with 256 grey levels per pixel

Examples

Source Code

results

Face Recognition using deep NN

Face Recognition using deep NN• First 6 images of each subject are used

to train the network

• Multiclass classification

Source code

Autoencoders

Multiclass classification results

Machine learning

Machine learningFeature presentation

Convolution

Alexnet

“We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes”

Spoof fingerprint detection using Convolutional Neural Network

Dataset• 4 Different fingerprint sensors • 5 different spoof fingerprint

materials

Source code

Result

Thank you