IKEA Harajuku Tokyo
Award winning AR shopping assistant for IKEA's Harajuku store, made with Unity.
Overview
The IKEA Harajuku app is a brand-new experience combining augmented reality, computer vision, and machine learning into one intuitive, meaningful design. Built for customers of IKEA's Harajuku flagship store, it was an experimental project exploring how cutting-edge technology can help with an everyday activity like shopping.
Our goal was a simple point-and-discover experience where users can shop and add an item to their cart simply by pointing the app at a product. The app recognizes the product, provides a link, and lets the customer add it to their cart and check out — all while browsing the aisles.
My Role
I was brought onto the project as a Unity expert, tasked with evolving and integrating the backend with the frontend designs. Working alongside the UI/UX team, we built a unique, engaging experience that intuitively told users what they needed to know, and when.
I had an amazing time building the app, and it's earned wonderful praise from its users — the work went on to be recognized with an iF Design Award and featured in iF's winner ranking for its user-interface design. You can download and try it yourself using the App Store link below.
Machine Learning - AI Integrations
A core challenge was teaching the app to recognize IKEA's products on sight. We developed AI models that could identify a subset of the IKEA product portfolio. We focused primarily on the items that were available in the Harajuku store.
To train them, we used Blender to generate simulated scenarios, building custom pipelines that produced the varied, labeled imagery the models needed. This approach pushed every model above 80% recognition accuracy across the Harajuku store's product range.
iF Design Award 2021
User Interface (UI) · Digital Media Interfaces
IKEA Harajuku was honored by the iF Design Award in two categories and featured in its winner ranking — read the full write-up.