As of now, the team has a functional web application that tracks swimmers’ splits and strokes. Moreover, it presents the additional functionality of delimiting stroke times on a continuous graph for users to see. Finally, our web-application also suggests workouts to swimmers, and allows clients to interface with it by picking a stroke and distance.

For our stroke classification, the ML portion of the project, we decided to opt for background subtraction + more pattern recognition to determine the appropriate stroke of our swimmer. The team found this approach more effective than the earlier approach of limb detection + relative coordinate input to a NN.

All in all, there is confidence that a decent capstone project will ensue, and we are looking forward to meet those ends…

This week is our final push to put together our project before the demo. The team worked to collect better footage, fine-tuning the tracking, and trying new techniques for stroke classification. We haven’t quite hit our metrics we set at the beginning of the semester, but we are pleased with our effort and results despite obstacles encountered during the semester.

At this point, the greatest risk to our project is how we present it at the demo day. We won’t be making any significant changes to the project itself, so we want to focus on how we will show our work to the outside world.

This week I spent a lot of time putting final touches on our project. One piece that needed improvement was tracking, and Adithya and I went out to get better footage so that our tracker could be more consistent.

Mostly, however, I spent time working on stroke classification and putting things together for the final demo. I have spent a good chunk of time working with the team to extract meaningful features from our video footage, as well as trial and error with machine learning models for classification.

One more day of hard work and we’ll be through the finish line!

I spent most of this week putting our project together. This included integrating the different components together and coming up with a plan for the impending demo on Monday. In particular, most work was focused on our stroke classification algorithm and tuning our model to get a higher accuracy. Additionally, I made minor adjustments to our web app front end to improve the user interface.

For the upcoming week, I hope to have a smooth demo and a strong report to finish up the course.

I spent many hours this week working on finalizing the web app with Adithya. The entire process of capturing, uploading, viewing, and exporting workouts is complete with no known bugs. I feel that this part of our project looks really good!

After completing the web app, I turned my attention towards stroke classification. We chose to use separate footage from a perfectly aerial perspective for this purpose. The goal is to extract the shape of a swimmer’s body as best we can to train on that. We have currently achieved 56% accuracy.

This week my aim was to have an UI working from end-to-end. To that end, I focused on the web app with Jack and it is now working completely. Furthermore, we made some progress w.r.t. the stroke classification task and we now have some naive classification techniques which we plan to build upon.

In the upcoming final week, our plan is to tweak the classification and tracking algorithms as much as possible to have a well functioning project in the end.

This week, the team has shown quite a bit of progress on the web application. In particular, it has quite a bit of commendable functionalities such as graphing the splits of swimmers’, as well as suggesting workouts to a swimmer. Moreover, the user has the option of picking a workout, as well as a choosing stroke/distance on the web application.

Unfortunately, we have made little to no progress on the limb detection, and subsequent stroke classification. This is most likely a consequence of the fact that we have not set up OpenPose, the main limb detection library. It is very possible we may not finish the stroke classification and/or limb detection; however, we plan to document all attempts to perform this feat in the final report.

As our deadline comes closer, we have effectively finalized all we need to do to get a sufficient capstone up and running. First, we need to continue work on the web app. In particular, we need to garner data of the wall touches of respective swimmers while sending that data to the web application, which will be plotted on a front-end UI.

Second, we need to work on the actual stroke classification. For this task, we need to input stroke information about respective swimmers into a NN, and then classify one of four strokes. The challenge with this task is figuring out how to do the limb detection itself, which has proven to be very difficult up until this point. Once we have the limb detection, though,  we can proceed to inputting relative positions of joints to the NN, and training our model.