Seeking collaboration: AI-powered foot measurement & size recommendation using OpenCV (for online sports shoe brand)

Hi

I represent a sports shoe brand selling directly to customers online. In our sport, proper fit is extremely critical – even small deviations can affect performance and comfort. However, since customers can’t try on the shoes before buying, they often struggle to find the right size. This leads to long decision times, heavy customer service workload, and high return rates.

We believe a computer vision‑based solution can solve this, and we’re looking for technical collaborators to build it together.

The idea:
Develop a mobile app (iOS/Android) that uses OpenCV to accurately analyse a user’s foot from just a few smartphone photos. The app would extract key measurements:

  • Foot length & width

  • Girth (instep, forefoot, etc.)

  • Arch type and overall foot shape

These data would then be fed into a trained model that recommends the optimal shoe size for our specific product lines.

Why this could work now:
A few similar apps exist, but they suffer from low accuracy, overly complex user workflows, or unclear use cases – none are truly production‑ready or widely adopted. We believe the missing pieces are:

  • Higher precision in vision‑based foot measurement

  • A simple, intuitive photo‑taking process

  • Real‑world validation with actual shoe fit data

We already have a dataset of user feedback and size‑fit relationships for our shoes. Combined with a robust OpenCV‑based pipeline, we can quickly iterate and validate the solution on real customers.

What we are looking for:
A developer or small team capable of:

  • Building a cross‑platform app (Flutter/React Native + native OpenCV)

  • Implementing foot detection, landmark extraction, and measurement from images

  • Training and optimising the size recommendation model (classical ML or deep learning)

What we offer:

  • Product & business context, plus existing fit‑related data

  • Real‑world testing through our brand’s customer base

  • Equity (technical shares) in the project – we prefer a partnership rather than a one‑time contract

We are a Chinese brand, so Chinese‑speaking collaborators would be ideal (for easier communication), but we are absolutely open to anyone worldwide who shares the vision.

If this sounds interesting to you, please reach out at: felix.luo@pl-sports.com

Let’s build the first truly reliable foot‑sizing tool and fix online shoe fit for good.

Thank you!

Word of advice: don’t lock down technical decisions (“use OpenCV”) from the management/founder side. Let the engineers try a few approaches and make an empirical decision. If it’s going to be a mobile app anyways, the smart move would be to use each platform’s own AR library instead.

Yes, your advice is absolutely correct. I have very limited technical knowledge in this area and only have a rough idea. I happened to learn that OpenCV can make judgments automatically through vision by building models, so I thought OpenCV might achieve what we want. If there are better approaches, they are certainly very welcome.