Image Recognition with Deep Learning - From Prototype to Production

Explore the development of an image recognition system for detecting clothing characteristics using deep learning and neural networks in this conference talk. Learn about the journey from prototype to production, including model training, system development, and creating production-worthy applications. Discover how to access a Kinect from the web and interface with the image recognition system. Gain insights into implementing deep learning techniques in customer-facing applications through code examples in Java, Node.js, React, and Python. Follow the process of creating and training models, categorizing images, and moving from categorization to recommendation. Understand the challenges faced and solutions implemented in developing a clothing analyzer, from initial tests to improved versions.

Intro
What we've built
Deep Learning
Training your kid (or neural net)
Convolutional Neural Networks
How we trained our model
Creating the model
Training the model
Categorizing new images
Categorizing images
From categorization to recommendation
Lessons Learned
The idea...
Steps to take...
Putting it all together...
How was this made?
How was this made: data
Clothing Analyzer - Difficulties
Clothing Analyzer - First test...
Clothing analyzer - Version 1
Clothing analyzer -Solutions (V2)