The problem
Important details in posture and timing are easy to miss in real time—especially under fatigue.
03 / Projects
Another set of eyes.
StrokeScore AI began with a rower’s question: could ordinary video make the mechanics of a stroke easier to see and discuss?
SS
Main project / in development
StrokeScore AI
Markerless, video-based rowing analysis that explores stroke phases, joint positions, timing, and movement consistency.
Explore StrokeScore AI ↗From question to prototype
The question
Could pose estimation turn a side-on rowing clip into movement information a rower can actually understand?
The experiments
I explored joint tracking, stroke phases, angles, motion trails, and the differences between sculling and sweep.
What I built
A tool that places explainable movement overlays directly on a rower’s own video.
What I learned
A visualization can begin a better conversation, but a clean number is not the same as a complete answer.
Still unresolved
Camera angle, occlusion, individual bodies, and context all limit what a model can know.
Photo to add
Full-width StrokeScore AI screenshot showing a real side-on video with angle overlays.
Photo to add
Early notebook sketch, interface wireframe, or development note with Hugo’s annotations.
Photo to add
Short project demo video poster: one stroke, clear pose landmarks, no private athlete data.
Useful, with limits.
StrokeScore AI is an experiment in seeing movement more clearly. It is not a medical device, a finished professional coaching product, or a replacement for coaches, trainers, or clinicians. Technique decisions still belong to the people who know the athlete and the water.
Future project template
Next experiment
01The question I started with
02What I tried
03What I made
04What changed my mind
05What remains unfinished