Pinterest Visual Search

In 2014, I built infrastructure for image feature extraction and object detection at scale to power visual search at Pinterest. We published a paper at KDD 2015 describing how visual search worked at that time.

Visual Search at Pinterest

Yushi Jing, David Liu, Dmitry Kislyuk, Andrew Zhai, Jiajing Xu, Jeff Donahue, Sarah Tavel

Abstract: We demonstrate that, with the availability of distributed computation platforms such as Amazon Web Services and open-source tools, it is possible for a small engineering team to build, launch and maintain a cost-effective, large-scale visual search system. We also demonstrate, through a comprehensive set of live experiments at Pinterest, that content recommendation powered by visual search improves user engagement. By sharing our implementation details and learnings from launching a commercial visual search engine from scratch, we hope visual search becomes more widely incorporated into today’s commercial applications.

Paper download