Seenit
Using machine learning for fine-tuned video search in the cloud
Industry
Customer application
- Video crowdsourcing and editing application
Solution
Use case
- Recommendation engine
- User profile
- Media catalog
Product
Key features
Built from the ground up with Couchbase Server and Mobile, Seenit gives enterprises a highly innovative and collaborative video platform that’s easy to use. After crowdsourcing video content from employees, fans, and customers, companies can quickly sort through thousands of videos to find the perfect clips. The powerful combination of Couchbase’s N1QL and Full-Text Search on top of machine learning in the cloud allows users to filter by visual objects in the video, words or phrases in the audio, sentiment, and other key attributes.
“We already extensively use N1QL for our structured querying, but now with full-text search integrated into Couchbase, we can seamlessly search all this data and derive relevance-based intelligence using a single data platform.”
Dave Starling
CTO, Seenit
CHALLENGES
- From vast catalogs of videos, find short clips that meet specific requirements
- Evaluate, store, and search complex properties of video content
- Scale to accommodate a quickly growing business, including new features, large files, and massive amounts of data
OUTCOMES
- Full-Text Search allows sophisticated search on any combination of visual objects, words, and sentiments
- Video tags stored as JSON objects are searchable using wildcarding, fuzzy search, and Boolean search
- Couchbase scaled effortlessly, and a new search feature implementation was shortened from 12 weeks to 1 week
Seenit and Couchbase resources
