Image Recognition tool for MSU Internet Connection

Inspiration

Our inspiration was helping peers and future students who may have difficulty connecting their devices to the MSU student Wi-Fi.

What it does

Our solution involves the student taking a picture of the device that they would like to connect to the internet. They then receive detailed instructions on how to properly connect it based on what type of device it is.

How we built it

We built our project using Domain.com to register the domain "MSUWifi.tech". We used Wix to design the website that serves as the example UI. We worked with both Google Cloud and Google Teachable Machine to create a machine learning algorithm that could accurately recognize devices based on the user's uploaded image. Wix was unable to work with the necessary files to use the machine learning algorithm, so we made an example UX for photo uploading and recognition using Visual Studio Code, GitHub, and Netlify. Once the photo is recognized as a type of device, the user is shown detailed steps with pictures guiding them to find the MAC address for that device. Finally, the user is redirected to the MSUIT's DCHP registration website where they can enter their device's MAC address and connect it to the Wi-Fi.

Challenges we ran into

We were unable to integrate the separate pieces of our project together into one fluid user experience due to a couple of technical difficulties. We initially wanted to use Google Cloud as our machine learning solution because it is more powerful and more accurate, but once we were training it we learned that it was going to take hours to adequately train and we did not have the time available to wait for that process. Our second issue was that we were unable to properly integrate the Google Teachable Machine because uploading the necessary .bin file is not supported on Wix, and the Wix iFrame object did not properly display the website for the image detection.

Accomplishments that we're proud of

No member of our group had previously submitted any projects as a part of a Hackathon event and our primary goal was to make a simple project that we could present to gain experience. We were able to go far beyond that and work in areas that none of us had previously explored. As a group, we had no prior experience with machine learning, HTML code, or using Wix.com to make a website. We are proud of our project even given the flaws it still has and are excited to be presenting it.

What we learned

A lot of our time was spent working to connect different technologies with each other. This included API calls using Wix.com, making requests to the Google Cloud and Google Teachable Machine, and uploading and pulling from GitHub. Also given our lack of experience with the concepts previously listed, all progress made in any of those areas involved a lot of learning, trial, and errors that we had to work through and troubleshoot.

What's next for Image Recognition tool for MSU Internet Connection

If we were to pursue improvements in our project for public use, we would need to update our solution to work with more types of devices. This would include training the machine learning to recognize more images, uploading more guides with photos and descriptions, and ensuring that the various parts of our project were able to work together seamlessly.