Computer Vision

NAMI

DEC 2024

NAMI LAB

8 week Project - MDE Design Studio course - team of 3

Nami is a fitness product that helps elite swimmers optimise their performance with data, rather than guesswork.



Solution

NAMI provides swimmers with a two way solution to improve their performance:

  • NAMI Wearable
    A modular IMU-based device designed to collect precise swimming motion and movement data, adaptable to its placement location

  • NAMI App
    A digital interface that provides coaches and swimmers with performance insights and helps them identify areas for improvement

Computer Vision

We used the YOLO model version 11 to create pose estimation videos of underwater swimming footage to help swimmers identify areas of improvement in their posture. We then built upon the pose estimation output to create an abstract view of the swimmer and a chronographic representation for deeper analysis in the stroke pattern.





Sensor Data

We can combine this visual data with 3D acceleration data which can act as a pinpoint signal on what frame of the video the swimmer needs to analyse.