adidas is one of the most widely recognized sports companies in the world. As the home of sport and sneakers, the adidas app puts their users closer to the action with instant access to sneaker drops, seasonal fashion releases, athlete inspiration and more.

In an adidas retail store, consumers want to know if a shoe is available in their size. While adidas store associates have tablets to check stockroom inventory in real-time, the adidas app lets consumers check themselves.

The adidas app team had opted for traditional visual search to make it easy for the consumer to find a shoe without having to scan a barcode. However, if a consumer took the picture too close or too far away it could lead to a frustrating consumer experience with an inaccurate result and time wasted.

With the introduction of ML Kit’s Object Detection & Tracking API, the team was able to create a seamless and intuitive experience. After opening the adidas app on Android or iOS and hitting the ‘try it on’ button, the phone’s camera is activated and using this API, the phone seamlessly detects shoes in real-time and within seconds returns an image recognition match against hundreds of products.

ML Kit detects objects of interest in real time which in the case of adidas are sports apparel and footwear. Based on these detections, the adidas app prefetches image recognition results from a cloud-based image recognition service before the user even explicitly requests results for the object of interest. This greatly increases the speed with which users get results and leads to a seamless and pleasant visual search experience.

“Using ML Kit’s Object Detection & Tracking API, we’ve not only been able to greatly improve the speed and accuracy of capturing what the user is looking for, but we’ve also able to provide them with a seamless consumer journey with ‘Bring it to me,’ said Chris Kettle, Product Lead for Digital Retail at adidas. “After the consumer scans the shoe, selects their size, an associate can bring out a pair anywhere in the store. Everyone loves this feature.".