Techniq

Techniq is a computer vision athlete analysis app that can be used for virtually every physical sport. Techniq takes professional footage the user inputs (i.e. Steph Curry shooting a basketball) and also takes the user’s own footage (i.e. themself shooting the basketball). Techniq can sync the two footages down to the frame and will then give the user frame-by-frame comparison of the user’s form against the professional. Techniq will give differences in form in degrees for all parts of the body.

Examples

Below you can find 2 video clips: One of me shooting the basketball, and one of Steph Curry shooting the basketball.

As evident by leaving the two videos playing, these video clips are not synchronized.

Running the program on these two inputs, the result generated is a series of side-by-side frames:

We rise at the same time and shoot at the same time. Differences in form now can be much more easily discerned.

Another example with the figure skating jump, the axel. Here are the initial clips:

And here is the output:

I (on left) am doing a double axel while Yuzuru Hanyu (on right, 2-time gold Olympic medalist) is doing a triple axel, which is one more rotation.

Notice even with significantly different camera angles, Techniq can still correctly synchronize.

Who is this for?

Techniq should be used by elite athletes and coaches alike for form training.

During day-to-day practice, athletes may not always have a coach to give them feedback. Currently, athletes rely either on purely repetitions or recording themselves and reviewing the footage. I believe Techniq solves that issue by giving instant high-level feedback to the individual athlete. And while this app obviously can’t completely replicate a coach, it can be a pivotal tool that speedens the learning feedback loop for the athlete.

Coaches can also greatly benefit from this. An athlete and a coach have limited time together, and that time is wasted if the two of them stand around for the coach to manually sync the videos. Additionally, the app may even spot faults in form that even the coach could not detect in real-time.

Why did we make Techniq?

Techniq was a product of a HackGT8, by Team STAC. STAC was composed of Charlie Liu, Amal Chaudry, Tina Nguyen, and Sheza Chaudhry. I came up with the idea, as I was an elite figure skater in high school and had experience using Google MediaPipe and OpenCV. I had a personal need for the app, and even after HackGT8, have continued to work on it independently.

Future

I’m currently actively building out the app, bringing it to mobile in the near future. The current project is currently showing great promise, having tested with a wide range of athletes to validate my assumptions and gather feedback to enhance the working model.

Specifically, current work is underway to transition the app to React Native, leveraging MediaPipe’s JavaScript library as the basis, with a backend database and frontend app that has account and video storage capabilities. I’m also closely working with a designer to plan out the the app’s overall design.

Be sure to check back in a few months on my progress!