team photo

Team 28

Team Members

Faculty Advisor

Jamey Calabrese
Jay Chandran
Everett Han
Yuwen Jin
Adam Veilleux

Yufeng Wu


UConn School of Engineering

sponsored by
Sponsor Image Not Available

The purpose of this project was to abstract away some of the more technical aspects of configuring a neural network for the purposes of image classification. Specifically, our intentions were to create a platform where users could create and manage image datasets, create models from those datasets, and classify images against them. Our final product did not fully realize those goals, but we were able to prove that a scalable, distributed, classification system is possible. This was accomplished through the development of three pieces of software. 1. The Worker - A thin python program which connects to the data service, to be discussed, in order to fulfill any user submitted classifications. There may be any number of these instances enabled at any moment. No request will be assigned more than one worker assuming the currently assigned worker produces a result within 72 hours. Github 2. The Data Service. - The core component of the platform. This is the centralized storage location for all of the image classification submissions. There are a collection of operations made available by this service to go about submitting, classifying, and viewing the results of a classification against a specific image. Github 3. The React Webapp - The user facing web application where images may be uploaded and results obtained. Github