Improving metadata tagging accuracy with AI and Computer Vision

• Background

  A leading Digital Asset Management Company used by 20 of the FTSE 100 companies and over 500 organizations around the world wanted to improve their metadata tagging accuracy. The company's platform provides their enterprise customers with the ability to manage their media libraries at scale.

• Information

  Use Case

  Metadata generation

  Industry

  Digital asset management

  Client

  Digital Asset Management Company

Challenge

As one of the industry’s innovators, the company needed to find a faster way to make it easier for their customers to search content and related assets. They understood that building artificial intelligence into their DAM platform and leveraging image auto-tagging would create a better product and user experience.

Solution

The company launched two pilot apps incorporating Clarifai’s computer vision platform to auto-tag images. The first app was built into the company’s interactive’s flagship product, so tags could be added automatically to assets as they are uploaded to the platform.

The second app was a standalone product with its own user interface which was designed to be used with any DAM solution. It allowed clients to change, accept, or reject automatically suggested tags. With these two pilots in place, the company could deliver added value to their clients and save time traditionally spent manually adding keywords to images.

Results

Clarifai’s General Model helped the company recognize over 11,000 general concepts, which far exceed the amount needed by the company’s clients to extract value from their results. In their platform, these general tags were automatically added via the Clarifai API as users upload images. For example, the general tags for “outdoors” and “indoors” were very meaningful for a travel and tourism client. 

• "Clarifai's AI lifecycle platform helped and its computer vision technology helped us dramatically improve our ability to accurately tag our assets with better descriptions. Their AI-assisted solution also helped us speed the process over 10x faster over what any team of human annotators could do."

  CO-FOUNDER