Scale Image Recognition API

API for image recognition, powered by humans

Reviews

Discussion

You need to become a Contributor to join the discussion - Find out how.
Alexandr Wang
Maker
@alexandrw · Cofounder/CEO at Scale API
Hey PH! Excited to be on here again. We made a demo app to show the power of our API: https://annotate-demo.scaleapi.com/ (and it's wicked easy to deploy your own version and build on top of it: https://github.com/scaleapi/scal...) Recognizing images is one of the things that humans are undeniably better than computers at. Our mission at Scale is to bring human intelligence to every software application, and the bounding box API is a huge step. I'm here to answer any of your questions! Really excited to see what everyone builds 😁
Brandon Foo@foolywk · Co-founder @PolymailApp
Congratulations! Awesome work, team! 🎉
Alexandr Wang
Maker
@alexandrw · Cofounder/CEO at Scale API
Thanks @foolywk!
Vishal JoshiHiring@vishaljoshi · co-founder at http://withjoy.com
Congratulations Alex & Lucy. This is looking great. We hope to use this someday.
Alexandr Wang
Maker
@alexandrw · Cofounder/CEO at Scale API
@vishaljoshi Thanks Vishal! We would love to work with Joy—just tell us when :)
Greg 'X' Willis@greg_x_willis · Founder & Developer @ Taste Buff
I'm going to play with the Image Cropping side of this API. I have a few good use cases for it. Beyond recognizing a logo, etc. how well those it hold up if you want too focus on something in a image with numerous possible (i.e. The Painting on the Sistine Chapel) focal points?
Alexandr Wang
Maker
@alexandrw · Cofounder/CEO at Scale API
@greg_x_willis In general, the goal should be unambiguous instructions for any human. The instructions should have it be the case where most people would be able to identify it correctly without any additional context.
Jan Ahrend@janahrend · PhD Student, University of Oxford
Can this be used to build a training set for machine learning?
Alexandr Wang
Maker
@alexandrw · Cofounder/CEO at Scale API
@janahrend Yup for sure! That was probably one of the first things it was used for—building training sets for image recognition ML algorithms.