Inspiration
- Observing newcomers to the gym struggling with poor technique, coupled with limited staff unable to give exclusive attention to each student, inspired our solution. In addition, we realized that some veteran gym users who had never been corrected were still showing poor form, which negatively affected their results.
What it does
- On the SmartLift platform, you can upload a video of yourself performing an exercise. Using the Gemini 1 pro, the video is analyzed to assess your posture, stability, range of motion, and technique. The platform offers coaching tips and recommended exercises to help improve your performance.
How we built it
- We used Astro as our web technology and Gemini 1 pro as the artificial intelligence model. We utilized the MUX platform for video uploading and playback on the web.
Challenges we ran into
- Originally, we wanted to evaluate technique over time and pinpoint parts of the video where the technique was good and where it was lacking. This would have been much more valuable for the end-user. However, we found that Gemini still suffers from significant hallucination issues and time-related inaccuracies for SmartLift's use case. Thus, we pivoted to a simpler idea, providing a general analysis instead.
Accomplishments that we're proud of
- Successfully integrating all components of the project and delivering a functional prototype.
What we learned
- Before initiating the development of an idea, it's crucial to conduct necessary validations to ascertain the viability with current technology.
What's next for SmartLift
- We are looking forward to improvements in Gemini's video capabilities to develop a product that can truly add value to the fitness industry.
Built With
- astro
- mux
- tailwind
- typescript
- vertex
Log in or sign up for Devpost to join the conversation.