This image is currently processing. Please be patient...

fashionksu

KSU FashionTech Hacakthon

Our domain name is stoh.io , which is a play on words because we are all from Stow, Ohio!

Be able to try on an entire catalog without having to use a dressing room, and have it tailored to your likes!

We built a Magic Mirror (google: raspberry pi magic mirror), and hooked it up to a Microsoft Kinect V1. The Kinect provides joint and frame data, on top of which we overlay an article of clothing. The image follows your frame as well as the Kinect will allow.

The system allows you to browse a catalog of clothing, with different gestures corresponding to different commands. If you like an article of clothing, a profile is built for you and a recommendation system (using collaborative filtering) suggests other clothing you may be interested in.

Built With

Submitted to

Fashion / Tech Hackathon 2016
- Winner Best Creation of a New Technology

Created by

I created the collaborative filtering recommendation system using SQL and PHP. Did work on the backend of the C# program to more easily do HTTP requests and JSON operations.

Istvan Gates
I set up the interface between the Kinect and the program with c#, including drawing the user's skeleton onto the screen, scaling the joint and bone data to match with the reflection, and setting up the bounding box to place the clothes on the user. I also set up interfaces to load clothing data from the web dynamically and putting the clothing onto the screen.

Cameron Sinko
I created a RESTful API for this project to be able to interact with the mirror and a MySQL server via PHP, all being routed through the Kinect C# app. I also setup the database to use Amazon AWS.

I added Pinterest integration. Login authorization with Pinterest and researched auto pinning your selections from the mirror to Pinterest.

MATT GATES
I worked on the layout and cleanup of the front end user interface in xaml and c# to coordinate the data coming and going through the interface.

Chris Paxton