Computer Vision – Foot Measurement App
Mobile native (Android & iOS) application for foot size measurement, which uses embedded computer vision (CV) model for detections and measurements for further shoe suggestions according to the processing results.
I must say that the App looks great! Nice work of you and the team!I must say that the App looks great! Nice work of you and the team!
Aim of the Project
- Embedded accurate solution for foot size measurement
- Cross-platform mobile application
- Measurement accuracy ±2.5mm*
- Processing speed: ~1 sec
- Native iOS & Android apps
- For ML/CV: TensorFlow / TensorFlow Lite / OpenCV / Python / C++.
- For Android: Kotlin, JNI.
- For iOS:
The understanding of the market of such applications was really important. We did research of similar apps that were available in the App Store and Google Play in June – July of 2020th. We found out that our apps were the most accurate on the market.
The UI/UX design was also developed by APRO:
What Apps Can Do
● Measure foot size in a really accurate, fast, and convenient way and automatic suggestions for the shoes of optimal sizes.
● ML/CV model works 100% locally without the need for an Internet connection. The Internet is needed for catalogues and map options only.
● Automatic measurements conversion in different sizing systems: EU, UK, MONDO.
● The applications can work on both smartphones and tablets on iOS and Android (ML/CV model is cross-platform too).
● The catalogue has the specially developed 360-degree shoe view.
● Google Maps Integration with the suggestions of nearest brand stores.
Main Sources of Benefits
● The ability to suggest “perfect” shoes for brand customers. It means, customers will have unique user experience and much less worry about issues with shoe sizes.
● Significant reduction in shoe returns for online orders. And a reduction in associated costs (logistics, transportation, warehousing, etc).
● Unique applications that can really improve customers marketing activities and bring “technology” to the brand image.