Case study

Suits U

Filed under:



Inspired by his own experience of having to choose from a vast range of glasses frames, with little confidence in his ability to select the pair best suited to him, Dr Lewis Griffin, Senior Lecturer in the Department of Computer Science and Joint Director of the CoMPLEX MRes, has applied his own sector expertise in Computer Vision to develop the Suits U software.


The Suits U prototype comprises two main components. To the user, it is a fun and simple interactive front end on a website or in-store iPad, in which a sequence of frame choices are presented, superimposed on a photo of the user’s face. Taking into account the user’s selections, the system then presents the most suitable choice, selected from the retailer’s entire database (typically upward of 1000 pairs).


However, this seemingly simple process belies a novel intelligent core, which includes complex machine vision algorithms acting together with substantial proprietary embedded knowledge regarding the factors influencing personal choice. This knowledge is experimentally gathered using machine learning techniques to determine the complex interaction between the aesthetics of glasses, face shape, personal taste and fashion trends.


Developed under London Development Agency Proof of Concept funding awarded by UCLB, Dr Griffin and Marina Santilli, Senior Business Manager at UCLB have since demonstrated the prototype to several of the largest high street and online retailers, with consistently enthusiastic feedback. The problem identified in high street stores is that a greater choice of product has led to longer browsing and selection time, resulting in increased demand on sales staff. Meanwhile, online retailers face the challenge of improving the navigation experience where the inventory is mainly hidden within a list of category tabs. Suits U is expected to alleviate both of these problems.


A follow-on PoC award will take the patented prototype to an installable demonstrator product in early 2012, when it will be implemented under licence for first customer trials.



Print friendly version of Suits U