SlideShare a Scribd company logo

Pley Technical Architecture

•Download as PPTX, PDF•
•
Anurag Phadke
Anurag Phadke

This document discusses convolutional neural networks (CNNs) for image classification and recognition. It provides examples of using a CNN to determine if an image contains a human face or is a picture of George Clooney. It also discusses how CNNs are trained on images and can be used to classify images, such as finding a green LEGO flower brick in an image. The document outlines the architecture of a system using Laravel, AngularJS, Amazon Web Services, and other technologies for its website, warehouse, customer service, and mobile apps.

Related slideshows

Convolutional Neural Networks for Computer vision Applications
Convolutional Neural Networks for Computer vision ApplicationsConvolutional Neural Networks for Computer vision Applications
Convolutional Neural Networks for Computer vision Applications
 
Pioneer a Strategic Change in Content Organization with Plone
Pioneer a Strategic Change in Content Organization with PlonePioneer a Strategic Change in Content Organization with Plone
Pioneer a Strategic Change in Content Organization with Plone
 
Aiden kelly y1 gd engine_terminology
Aiden kelly y1 gd engine_terminologyAiden kelly y1 gd engine_terminology
Aiden kelly y1 gd engine_terminology
 
1 of 19
Pley Technical Architecture
Pley Convolutional Neural Network for Images Used by: -eidetic.
Machine learning is used for two things: 1. Search / Classification 2. Recognition / Matching
Key Question Does this image contain any human face?
Key Question Is this a picture of George Clooney?
First Principles Show images of correct “object” multiple times Show images of incorrect “object” multiple times
Traditional Machine Learning • Lots of training data • Carefully crafted features Makes the solution exponentially difficult
Classifier to find a green LEGO flower brick
Eidetic - CNN
Classifier to find a green LEGO flower brick
Demo http://54.197.37.188:34448/models/images/classification/classify_one?job_id=20160225- 194225-0741
Demo http://54.197.37.188:34448/models/images/classification/classify_one?job_id=20160225- 194225-0741
LEGO mixed bricks
Architecture Key Components Frontend: - AngularJS 1.2.26 - Plugins installed via Bower Backend: - Laravel 4.1 (PHP 5.5.18) - Plugins installed via Composer - Optimized for REST APIs Email: - Amazon SES + InfusionSoft / ExactTarget Deployment: - Gulp Infrastructure: - Amazon EC2 Payments: - Braintree Shipping Logistics: - EasyPost Source Code: - GitHub Alerts: - PagerDuty + Pingdom
Pley Backend - REST APIs Used by: -website. -warehouse. -customer service. -iOS (and other apps)
Database Amazon RDS (MySQL) CloudFlare (DNS) REST API (Laravel 4.1 / PHP 5.5.18) Apache HTTP Server 2.x Backend Load Balancer (Amazon) Horizontal Scaling Horizontal Scaling
Pley Frontend used by: -website. -warehouse. -customer service.
Browser (Client side) Apache HTTP Server 2.x HTML5 CSS3 jQuery Bootstrap Media Queries Frontend Load Balancer (Amazon) Content Delivery Network (CloudFront + CloudFlare) Horizontal Scaling
Thank You :-) Questions? anuragp@pley.com

More Related Content

Pley Technical Architecture