Cortex API, a general purpose brain-aligned embedding. Image In. Vector Out.
More Features
Brain Alignment relinquishes the need for billions of images
Not requiring training data helps us leverage and respect your privacy
Eye-blinking speed for computation of embeddings
Embeddings can be used for any computer vision task beyond search
The visual search based on Cortex allows users to efficiently search for products using images from their mobile phones or desktop.
"Brain Aligned" refers to designing and developing algorithms or models that are inspired by the principles and functions of the human brain.
Embedding models do not need any extra training data in general if they manage to tile and learn the whole image space. The key advantage of ours (and many other) embedding models is that we do not need extra training data since we are outputting vectors rather than labels. Overall, the vector space we learn is better than most other models as can be verified through the Brain-Score competition.
You can do many things with Cortex, like Visual search for your site, Compare images, and improve your product recommendation are just some of the many use cases of Cortex.
You can understand more in the Cortex Documentation.
