Some details
About the project
This app recommends the right apparel size based on a 3D model generated from user photos and body measurements.
The process is simple - users provide only a couple of front and side photos of themselves, as well as their height and weight measurements. A 3D model is generated with 99.5% accuracy and is ready to use when choosing clothes. Based on data provided by our partners we
suggest the right apparel size for each user.
How we helped
The solution was complex in terms of implemented technologies - neural network and custom algorithms, API, native iOS and Android apps. We performed a very detailed competitive research and identified a product niche. The team not only delivered the product but also provided the strategic vision for it. Meanwhile, we managed to achieve the following:
- Reach high levels of accuracy of human body measurements
- Reach high levels of 3D model accuracy and presentation
- Create and train a neural network
- Develop a well-documented API that could be used both internally and externally
The team created intellectual property in the domains of background subtraction, human body registration, human shape fitting, and measurement extraction.
General info
- Technologies used: Python, Django, Angular, SWIFT, Java, C++, Amazon Web Services
- Domain: 3D, fashion
- Team: 6 engineers, 1 QA, 1 Architect, and 1 Product Manager
- Duration: 18 months