Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
La reconnaissance d’images par réseau de neurones
Search
walid chergui
February 17, 2017
Programming
0
38
La reconnaissance d’images par réseau de neurones
walid chergui
February 17, 2017
Tweet
Share
More Decks by walid chergui
See All by walid chergui
Ansible, Infrastructure as code - VoxxedDays Luxembourg 2016
oraclewalid
0
160
Other Decks in Programming
See All in Programming
PipeCDのプラグイン化で目指すところ
warashi
1
260
ISUCON研修おかわり会 講義スライド
arfes0e2b3c
1
410
XP, Testing and ninja testing
m_seki
3
230
AWS CDKの推しポイント 〜CloudFormationと比較してみた〜
akihisaikeda
3
320
地方に住むエンジニアの残酷な現実とキャリア論
ichimichi
5
1.5k
設計やレビューに悩んでいるPHPerに贈る、クリーンなオブジェクト設計の指針たち
panda_program
6
1.9k
Blazing Fast UI Development with Compose Hot Reload (droidcon New York 2025)
zsmb
1
280
ruby.wasmで多人数リアルタイム通信ゲームを作ろう
lnit
3
430
datadog dash 2025 LLM observability for reliability and stability
ivry_presentationmaterials
0
450
技術同人誌をMCP Serverにしてみた
74th
1
620
C++20 射影変換
faithandbrave
0
560
今ならAmazon ECSのサービス間通信をどう選ぶか / Selection of ECS Interservice Communication 2025
tkikuc
21
3.9k
Featured
See All Featured
[RailsConf 2023 Opening Keynote] The Magic of Rails
eileencodes
29
9.5k
Evolution of real-time – Irina Nazarova, EuRuKo, 2024
irinanazarova
8
810
Into the Great Unknown - MozCon
thekraken
39
1.9k
Fight the Zombie Pattern Library - RWD Summit 2016
marcelosomers
233
17k
Agile that works and the tools we love
rasmusluckow
329
21k
How to Think Like a Performance Engineer
csswizardry
24
1.7k
Making Projects Easy
brettharned
116
6.3k
Let's Do A Bunch of Simple Stuff to Make Websites Faster
chriscoyier
507
140k
A Modern Web Designer's Workflow
chriscoyier
694
190k
Why Our Code Smells
bkeepers
PRO
337
57k
How STYLIGHT went responsive
nonsquared
100
5.6k
XXLCSS - How to scale CSS and keep your sanity
sugarenia
248
1.3M
Transcript
La reconnaissance d’images par réseau de neurones Walid Chergui #SnowCamp
2017
Sommaire Encodage Convolution Réseau de neurone simple Réseau de neurone
de convolution LeNet Frameworks & code ImagNet
Moi Développeur Scala et Java FP Machine learning Les architectures
distribuées DevOps
Les images - Matrice
Les images RGB
Convolution Source : Wikipedia
Convolution
Réseaux de neurones Convolutional Neural Networks (CNN) Recurrent Neural Networks
(RNN) Autoencoders
CNN - types
Réseau de neurones
CNN-LeNet
Convolution (1*32*32) -> (16*32*32)
Relu (rectified linear unit )
Pooling (16*32*32) -> (16*16*16)
Fully Connected
Frameworks Caffe Torch Theano / Lasagne TensorFlow / Keras Deeplearning4j
CIFAR-10
Deeplearning4J CODE CODE CODE !!!
Depper ….
AlexNet (2012)
VGG Net (2014)
GoogLeNet (2015)
Pre-trained neural network TrainedModelHelper helper = new TrainedModelHelper(TrainedModels.VGG16); ComputationGraph vgg16
= helper.loadModel(); DataNormalization scaler = new VGG16ImagePreProcessor(); scaler.transform(image); INDArray[] output = vgg16.output(false,image); System.out.println(TrainedModels.VGG16.decodePredictions(output[0]));
Merci ! @walidchergui