The year has wrapped up! The Fantastic 5 delivered our final presentation of our project and what he have accomplished in a year. We also showed off our prototype to the general public and our liaisons.

We had a lot of fun and got to meet our liaison Nick Rich in person for the first time which was exciting!

It is sad to see my teammates go, but we for sure will keep in contact and hope everyone has a successful future post graduation.

Thank you to our coach Dr. Thomas for encouraging us throughout our project and thank you to our liaison Nick Rich for also giving us support and helping us with our project!

This is the last leg of the project. The team focused on perfecting the prototype and getting ready for peer review FDR. We got our main deliverables like the video, poster, and report draft finished this week.

We recorded the video with ourselves in it this time and adjusted the poster to fit all the feedback we got from our peers. Along with that, we have our demo of the model ready for showcasing at the final design report. We did our peer review this week as well and got helpful feedback. We altered our presentation based on this.

We’re excited to show off our final product and can’t wait to see how our liaison will find it to be.

This week the team worked on getting all of our deliverables for the final design report. We created a draft for the final poster and video and received helpful feedback on both. We will use that to adjust the deliverables before the final day.

Our initial video is with animation, but we will involve ourselves to add a personal touch. We also continued working on training the model with more data. Along with that, the team wrote some of the final design report.

This week, our team finished prepping and presented our prototype to various judges that gave us feedback. We ended up having a useful API demonstration to showcase how the process would look on the user side. We also had an AI prototype that shows the primary model based on limited training data. This prototype helped us a lot because we know to add certain things during the final presentation that we did not have during this one, due to the feedback. We will continue to improve on this prototype using fine tuning and hopefully have a graceful final product when the time comes. Overall, we enjoyed showing off our work to the judges!

This week the team focused on creating a prototype for the Prototype Inspection Day next week. We have created a presentation plan that is interactive and showcases all our work so far. There will be a demo that we worked on this week. This involved creating a UI that judges can interact with to see how our API flow is expected to act.

We are hoping to have an AI model trained in time, but that is dependent on hardware. So far, training takes time.

We are glad to be back from the break and went right into work. We had a very helpful review of our data from our liaison and received new data that is much easier to train a model with. Our liaison used AI to categorize the images for us, so the images now say whether they are the front or back view with 90% accuracy!

The team also created a story board draft for the video we will create as our final deliverable. Along with that we are continuing testing and completing our API portion, which has made substantial progress. Sara is currently working on a UI for it to demonstrate during our PID the API flow.

The team enjoyed spring break!

This week the team continued downloading some data we received. We looked through the data and tried to find ways to sort through it. The data does not have clear indication in the naming scheme if it is an image of the front or back of the jersey, which is a challenge as there are a lot of miscellaneous images that will not help to train our model, such as zoomed in photos of jersey tags. We are in talks with our liaison to figure out a way to categorize the images better.

We also started working on a Final Design Report outline for the final presentation since it is coming up soon. We got our QRB 2 feedback back and are going to implement any useful critiques for the next presentation day.

QRB 2 PRESENTATION

This week, the team presented the QRB 2 to a group of judges. We were given feedback on our project’s current state and any advice that could be helpful for development. These are some of the notes taken during the presentation:

• Was asked by the judges about what happened when we trained the model.
  • We responded by explaining that image data was given just this week and training will begin after all images are downloaded and ready to be pushed into the model.
• Asked about how large the dataset was?
  • We said that 80,000 data points were provided; this is not ideal, but we have our training model set up to update itself, which could help us to increase its accuracy.
• Was asked if it’s possible to have a jersey that’s set as a standard and we just add things to it.
  • We said this is not possible because all the jerseys aren’t equally alike.
• Asked if we have spent money on OpenAI?
  • We haven’t because using stable diffusion is fit to better depict clothing items compared to OpenAI.

Overall, the presentation went well, and we got very good feedback.

Next week, we will be reviewing our QRB 2 results and discussing more about developing the FDR.

This week, the team had the honor to listen to a guest lecture Milliken and Company on the inclusion and diversity in the workforce. This lecture helped us learn more about what its like integrating in the work life culture and the continuous discrimination that still occurs in the workplace. The insights gained from the lecture have inspired our team to actively advocate for inclusivity and equality within our own organization and beyond.

In our team meeting this week, we were successful in collecting all the attribute fields we need for model training. We also finalized our QRB2 PowerPoint slide and are preparing our speeches for next week.

Next week, we will present our QRB 2 and get feedback on our project’s current state.