May 29, 2017
In our spring series, Senior Theses: A Celebration, we take a moment in the interlude between thesis deadlines and graduation to appreciate the diverse, personal, and impactful work of seniors’ capstone research projects.
~~~~~~
Xuewei Ouyang is a senior in the Computer Science Department. For her thesis, she combined her passion for dance and her knowledge of coding to create an app called, ChoreoSpot. Here’s what she had to say about her work:
What is your thesis about?
In short, my thesis is about creating an app that takes a rehearsal video and, within various frames of the video, spots errors on the dancers’ bodies in comparison to the choreographer’s.
What kind of research did you need to do for this project?
Because my thesis wasn’t just technical, I also had to look at it from an entrepreneurial aspect. On the one side, I did research into the dance industry and the market; I looked for any existing app that already focused on correcting movement. My research also included looking into how to do pose estimation; I had to work with different algorithms and figure out what might function best for a dance video. Lastly, I did some research on how to create the app and how to put all these different components together.
ChoreoSpot looking for discrepancies between dancers’ movements!
What is your preferred environment for thesising?
When I worked on my thesis, I was usually in the library or my room depending on the hours. But I especially liked the library because it’s a quiet environment that can help me concentrate. With other work, I feel like I can study in the middle of Frist because I don’t have to be fully focused. But when I’m working on my thesis, there are no set guidelines or steps on what to do next. So I really need to be able to focus; otherwise, I’ll get sidetracked.
What’s the inspiration behind your thesis?
I actually started out with a different thesis topic last semester. I was planning on creating a kind of visual fashion calendar. But then I started working on a project for my computer vision class where my group and I were looking into what would happen if we applied pose estimation to dance videos. Because the project was a one-week Dean’s Date assignment, we ended up not getting very far with the results. But after working on it, I realized I was interested in pursuing the topic further. So I switched my thesis topic half-way through the year and decided to create an app that does a more accurate job of picking up mistakes than a person can.
Is there anything that you can’t thesis without?
I definitely need coffee. On a regular basis, I drink one cup a day. But when I’m really sleep deprived, I drink 3 to 4 cups.
Are there any songs you like to listen to while thesising?
I tend to listen to very chill, mellow songs with lyrics. Otherwise, I typically didn’t listen to music that much while thesising; I found that it didn’t really help me more or less.
Did you have an “Aha! moment” when writing your thesis ?
The biggest jump for me was getting my pose estimation to work on a sample image. Before I could make any progress, I had to download some code and get it to work on the dance videos. But for some reason, I couldn’t get it to function on my computer. So I had to get a lot of help from people in the COS department. When it finally worked, I realized, “Wow, my thesis is actually going to happen.”
I wish I had been able to work on this project since the beginning of the school year, just because I basically had half the time after switching topics. But if I were to go back, I think I would have worked faster. While I finished the main parts of my thesis that I wanted, I don’t think it’s a perfectly working prototype. And if I had had more time, I would have been able to add in more features that would have made it even cooler.
— Interview by Taylor Griffith, Social Sciences Correspondent