PerfectRep - Train smarter, lift better, progress faster!

Inspiration

One of the members of our team plays rugby and, when working out, he is used to performing exercises similar to the ones of the olympic weightlifting sport. With that, he started to engage more with that sport. That, aligned with a documentary we've watched about the Portuguese unicorn "Sword Health", lead to us thinking of applying computer vision to this niche sport.

What it does

The main functionality of our product, is to monitor, optimize and improve the lifts of our athletes using computer vision. This consists on an app, with 2 programs, one for the snatch and the other for the clean and jerk. The user records his lift using two cameras (one frontal and one from the side), and the app monitors the movement, giving feedback on it. After that, depending on the feedback, the app suggests exercises the user should do to improve his posture and performance. This is useful in the sense of individual asynchronous training for the athletes, but also for a better monitoring of remote training sessions by the coach.

How will we build it

In terms of the computer vision technology, we will implement python code and loads of videos of correct lifts, from people with different body types and lifting styles, from the internet. We'll also be using personal data to enhance specific user performance. The recordings will be done using 2 cameras.

Commercial Potential

The commercial potential of this idea resides mainly on the payment of the license to use the app, through a freemium system that begins with a certain number of free tries, and then offers the possibility to pay for unlimited lift analysis. We also want to establish a source of income by selling cameras (with the minimum specs for the app to run efficiently), individually or in packs of 2, as not everyone has access to 2 working filming devices.

Challenges we ran into

We had to research about computer vision technology, to understand the viability of our product and how it would work. Also we had to contact various people who are expert in this sport, to understand better its ecosystem in Portugal and assess if this is a realistic solution to solve a real problem in this field.

Accomplishments that we're proud of

We've found a scientific study which explored specifically the application of computer vision in olympic weightlifting, which supported and validated our idea. Also, we've spoken with experts on this sport: an athlete, a cross-fit coach (specialised in olympic movements) and the vice-president of the Portuguese Olympic Weightlifting Federation, whose testimonies helped us realising that this was an actual problem and to improve our idea.

What we learned

We learned more about computer vision technology and the olympic weightlifting sport. Also we cemented our ideation and networking skills, while working on this idea.