LOS ANGELES--(BUSINESS WIRE)--CloudSight (http://cloudsight.ai/), a visual cognition technology, announces today it has secured two patents to expand what its advanced object captioning technology can offer merchants and companies. Now with two new and protected patents, CloudSight’s object-recognizing technology improves its ability to better recognize images with a human-like instinct and image sourcing.
CloudSight is able to offer its technology through two patent avenues. The first covers both image processing systems and image processing methods. In these methods, an image is routed to both an automated system and to a human reviewer. The automated system produces one or more image tags for the image, along with a measure of confidence that the tag(s) correctly characterizes the image; the human reviewer also returns one or more image tags. Image processing systems of the invention include, amongst other things, logic that receives a manual review of an image from a human reviewer and an automated review of the image and provides image tags from both reviewers to an advertising system. The system also includes logic that receives, from the advertising system, an identifier for an advertisement and in turn provides the identifier to the source of the image.
The second patent directs CloudSight’s methods for upgrading image reviews. Within these methods an image is received and posted to an automatic system that generates image tags. Those tags are then received and an upgrade request is automatically generated based on one or more of those image tags. Once an upgrade request is generated, a human reviewer is then selected, and the image is posted to that human reviewer allowing for multiple image tags.
“When we first created CloudSight we sought to create a system that replicates human-like vision and language,” said Brad Folkens, CTO and Co-Founder of CloudSight. “Now, our protected technology gives computers and users an easier and upgraded integration.”
CloudSight is a Los Angeles-based tech company founded in 2012 by two computer scientists and best friends, Brad Folkens and Dominik Mazur. With a growing database of over 400 million images consisting of high quality metadata and annotations, CloudSight utilizes real world data curated over the last four years to train deep learning neural networks to automatically caption images. The startup is backed by $8 million in funding, has over 30 patents pending worldwide, and counts companies such as Kohl’s, P&G, Belk, and many other top online destinations and Fortune 500s as customers. Its mission is to build a visual brain for AI for businesses around the world, so computers can have full image cognition.