Imagine a world where all visual content is instantly shoppable and available to virtually try on before purchasing. The field of Computer Vision, which aims to help machines understand visual data as a human would, is rapidly moving toward making that—and more—a reality, dramatically changing the way we shop and interact with the world around us.
eBay recently launched two new computer vision products that take us one step closer to that future: Find It On eBay and Image Search. These features allow shoppers to use pictures instead of words to search eBay’s 1.1 billion listings. Find It On eBay and Image Search are only the beginning. As shoppers continue to search with pictures, Find It On eBay and Image Search will get smarter, learn and improve on the results that are surfaced.
These features are just some of the ways eBay is taking the shopping experience to the next level with image recognition technology. eBay is using image recognition to help shoppers easily find what they’re looking for when using eBay ShopBot for Facebook Messenger; and, when shopping for fashion items, eBay’s Shop the Look helps buyers easily browse for items that make up iconic fashion looks.
To coincide with the launch of Find It On eBay and Image Search, I had the opportunity to attend the International Conference on Computer Vision (ICCV) along with eBay employees from our New York, Tel Aviv and San Francisco offices. The conference brings together researchers and industry leaders in the field of Computer Vision from around the globe to present state of the art work in the field, creating a rich environment for networking, recruiting, learning and, above all, inspiration. eBay was one of the ICCV sponsors, helping get the word out about how we are using Computer Vision to transform the shopping experience on eBay. With over 1.1 billion listings containing real world images, eBay has some of the richest image data around. This data provides our researchers with a veritable playground for developing computer vision products.
As a product manager for Find It On eBay and Image Search, keeping up with the latest research in Computer Vision is important for developing a forward-thinking product strategy underpinned by state of the art technology. This year at ICCV, one of the hottest topics was Generative Adversarial Networks (GANs). These models learn to create new content that is similar to the data they were trained on. For example, a model trained using images of outfits could generate images of completely new outfits. In addition to being mind-blowingly cool from a research standpoint, the applications for GANs in ecommerce products are endless and creating a ton of buzz in the space.
One exciting application is for fashion, which was overall a popular vertical at this year’s conference. eBay participated in a fashion focused workshop where Robinson Piramuthu, eBay’s Chief Scientist for Computer Vision, Thomas Balestri from Markable (a Friends of eBay startup) and a host of other top tier researchers from academia and industry presented on the unique challenges associated with applying Computer Vision research to fashion and recent advances in the field. Topics ranged from visual search and retrieval in video to personalizing outfit recommendations based on your own closet and provided serious inspiration for future applications at eBay.
The energy at the conference coupled with the incredible research being presented reinforced to me that this is a space retail companies must invest in now or risk being left behind. Computer Vision can be so much more than upping your selfie game with dog ears and flower crowns. The research taking place right now at universities and companies around the world is paving the way for the future of shopping – a future that eBay is sure to be a part of and one that I’m particularly excited for.
Mary Titus is a Product Manager at eBay in New York City, where she focuses on building innovative experiences powered by computer vision and machine learning. She is a huge fan of the theater, and when not in the office can often be found catching the latest Broadway show.