Read more here
The eBay Motors app has been in beta for a few months, and version 1.0 is now live for Android and iOS.
The new incarnation sports a number of notable automated features to help people sell their vehicles, including an image-categorization tool that groups photos by type, which was built using Google’s AutoML Vision Edge tool for Firebase. The app has also been rebuilt from the ground up using Flutter, Google’s open source platform for developing cross-platform apps based on the same codebase.
For background, eBay first launched a dedicated mobile app for cars back in 2011, but ended up pulling the app — alongside three other standalone apps — in 2015 as it consolidated its services under the main eBay platform. Fast forward to 2019, and eBay Motors has risen from the dead with a renewed focus on helping individuals compete with professional dealers with as little friction as possible.
It’s worth noting that eBay is no stranger to computer vision — the main eBay app has offered visual search tools for a few years. The eBay Motors app itself offered a visual search tool as far back as 2012, allowing users to snap a photo of a car in the real world to find similar vehicles on eBay.
But with advances in technology and edge computing, the 2019 incarnation of the app is a different beast, with all the image processing now taking place on the phones themselves. This shows how far we’ve come in terms of the power of off-the-shelf solutions such as AutoML Vision Edge. According to eBay, its algorithm was created by a single engineer in a matter of days, using just a few hundred photos from existing eBay listings to train the model.
Read more here
Join us February 26-27 for the Property Portal Watch Conference Bangkok 2020.