The general tagging update for Visual Recognition we announced last month saw users reporting significantly higher accuracies for general tagging due to an active vocabulary that was 2.5 times larger than before. That update is part of a series of ongoing developments we’re making to both improve existing general tagging capabilities and introduce new ones.
Seeing in color
A new capability being introduced to Visual Recognition today is color tagging. While Watson has already been able to detect color, it will now return the top colors it sees in each image as response tags, each accompanied by a classification score. The new capability allows users to quickly assess the dominant color schemes within an image and turn these into actionable insights.
IBM Watson partnered with iconic Melbourne Couture designer JASONGRECH to launch a custom collection that debuted at the 2016 Melbourne Spring Fashion Week. Using the Watson Visual Recognition API, company founder Jason Grech was able to not only analyze, but also predict trends in color from ten years of fashion runway images. Read more about the use case here.
All of these changes happen on our end, so you should see the color tags automatically appearing in your results. Combined with the pricing changes we announced last month, it’s now easier than ever to start building with Visual Recognition.
Need help or have questions?
We’re excited to see what you build with Watson Visual Recognition, and we’re happy to help you along the way. Try the general tagging feature and share any questions or comments you have on our developerWorks forums.