Crowdsourced Video Classification

Javascript
Web Applications
SQL
Node.js
Website
Software Engineering
Full Stack

Web application to crowdsource the effort to associate emotion labels with video clips. This application was developed for the 2019-2020 Oregon State University Computer Science Capstone. Xandr, an AT&T company, plans to create a system to place ads relevant to the main content of online video or TV broadcasts. First, they will train their system to recognize emotion in video. For this they will use a technique called "machine learning" which mimics the way humans learn by training the computer on a large number of examples. This is like how babies learn to talk after years of exposure hearing other people speak. So in order to train their system, they will need a large number of videos that have already been classified according to their content. Labeling so many videos would take too long for them to do alone, so they would like to crowdsource their effort. Crowdsourcing means they will ask for help from the general public in accomplishing their immense task. To facilitate the crowdsourced task, our group built a website allowing visitors to participate. Jared Beale worked on the Database API, Sam Young worked on the web pages, and Connner Maddalozzo contributed to each. For more information, please see the artifacts included below.

0 Lifts 

Artifacts

Name Description
Web Application The product itself, deployed on Heroku.   Link
GitHub Link The GitHub repository where our code is located.   Link
Expo Poster The poster we would have shown at the Engineering Expo.   Download
Showcase Video This video showcases our project in lieu of the physical Engineering Expo. It includes a high level description of the product and its usefulness, as well as a brief description of our solutions, then a walk through of the product.   Link