MIRO

Over the last years e-commerce has grown steadily and has become an indispensable tool for retail stores around the world. Nevertheless there is a substantial portion of consumers who still are apprehensive with online shopping. These concerns increase in relation to clothes and shoes because consumers are not sure if the product will fit. With this problem in mind, Visian Systems accepted the challenge to develop an application that will assist both virtual and retail stores. Using smartphones and tablets cameras the application will accurately measure the feet of customers and determine the appropriate shoe’s size to reduce product return rates.

Early development and machine learning

Due to the fact the project involves pattern recognition, our team organized itself so the initial app design and the training of the algorithm for recognition of feet photos occurred simultaneously to ensure the early development was not delayed.

Communication was essential to keep our team integrated with the entire development process and we managed to fast produce a prototype of the application that presented the basic functioning for testing.

Testing and optimizing the app 

The app entered its testing phase where we are verifying if there was any bug in the machine learning process while also working on modifications that will optimize Miro’s functions. 

Visian Systems’ commitment is to deliver the best possible user experience, which has been one of the main company’s priority in all our projects, and to also develop an app with intuitive design and easy interaction that will meet the needs of the customer with a practical and fast running solution.

Technology & tools 

To develop Miro our team used the React Native library which provides support for mobile applications development in JavaScript. We also made a few integrations with the Python-based Django framework.

For algorithm training we chose to work with two other libraries OpenCV and Dlib, both written in C++, since they are great for developing applications which utilizes image processing and machine learning.