Rakuten Taiwan launches visual search for fashion e-commerce

Instead of using keywords, Rakuten shoppers can now upload images to search for what they are looking to buy, thanks to its new visual search and recognition service powered by Singapore startup ViSenze.
Written by Ryan Huang, Contributor

E-commerce giant Rakuten has launched a visual search service on its Taiwan site to allow its shoppers to use images to find what they are looking for.

This allows users to search by uploading photographs of people on the streets or their favorite celebrities, which will generate matches and close recommendations. The feature is available on two sites: www.oshare.com.tw and www.rakuten.com.tw/event/funsearch.

Upload an image of something you fancy--shown in the top left, the visual search engine generates close matches. (credit: www.rakuten.com.tw/event/funsearch/)

According to Toru Shimada, senior executive officer and CEO of Rakuten Asia, it is the first to roll out such a feature in Taiwan and will be offering the service to other parts of Asia.

The Fashion Finder application is powered by startup ViSenze, using technology from the NUS-Tsinghua Extreme Search Centre (NExT)--a partnership between National University of Singapore (NUS) and Tsinghua University of China.

ViSenze's Fashion Finder gives recommendations matching your search.

According to ViSenze, its technology is different from competitors because it is more accurate and can handle more complex backgrounds.

In addition to Rakuten, ViSenze said its cloud based technology has also been deployed on Singapore-based shopping network Clozette.
Examples of similar technologies include Riya, Superfish, and Google Goggles. Visual search services have also been applied in supermarkets and museums.

Editorial standards